In Kennedy v. Louisiana, the Supreme Court noted that the problem of “unreliable, induced, and even imagined child testimony” creates a “special risk of wrongful execution in some child rape cases.” Indeed, empirical research has repeatedly demonstrated problems with accuracy in children’s accounts of their own experiences. Although the research and commentary in this area has focused on how allegations of child sexual abuse are addressed in the criminal justice system, these studies have much broader implications: every year, state officials conduct millions of interviews with children in the context of child welfare investigations. These investigations have serious consequences for families—for instance, they can lead to the placement of a child in foster care or the termination of parental rights. This article examines the reliability of child welfare determinations by looking at a subset of the information investigators consider: children’s statements about their past experiences. First, this article reviews empirical research on suggestibility, lie-telling behavior, and the capacity of adults to detect lies in children. The article then examines the impact of structural features in the child welfare system, and posits that these structural features do not facilitate the proper evaluation of child statements. The article concludes by proposing legal reforms to improve the reliability of child welfare determinations. Ultimately, this article aims to defend the proposition that caring deeply about children and their safety does not necessarily mean the child welfare system should rely on what children say.
Statistics reveal that millions of children each year are the subjects of child welfare investigations. While only a fraction of cases proceed to court, and only a minority of those cases result in the removal of a child from her home, nearly four hundred thousand children are in foster care in the United States. Although many more cases involving children are tried in family courts and dependency courts than in criminal courts, the stakes are just as high. In these cases, parents may be deprived of their fundamental right to parent, or even visit, their own children. Likewise, children may be deprived of a relationship with their parents and family.
As the millions of reports of maltreatment translate to hundreds of thousands of children in foster care, how is that information filtered? Are investigators and judges making accurate judgments about which cases present real risks to children? This paper deals with a subset of that question, the impact a child’s account of her past experiences may have on the accuracy of those judgments. This article examines how accurately, or inaccurately, the child welfare system is likely to evaluate statements that children make.
Children’s statements form an important part of a record of child maltreatment. As one treatise on the Legal Rights of Children notes, “[c]hildren are the most important witnesses to their own abuse or neglect.” Yet an enormous body of empirical research has documented problems with child testimony and fairness in the legal system. Despite the attention this science has received in the context of criminal sex abuse trials, child welfare cases continue to rely on the statements of children in similarly problematic ways.
This article will offer empirical psychological research to demonstrate some of the impediments to learning the truth about a child’s experiences. Part I of the article begins by describing the scope of the investigatory apparatus dedicated to child welfare cases. This section describes how a child’s statements, including statements about everyday activities, can be made a part of a child welfare case. Having established that children’s statements are a crucial part of these investigations, Part II examines empirical research on children’s statements. In particular, this part focuses on (1) the way in which certain forms of questioning may, unintentionally, elicit an untruthful response from a child (“suggestibility”) and (2) how factors such as the desire to avoid getting caught in a transgression may encourage a child to say something that is not true (“intentional lying”). Part III reviews research regarding the ability of adults to detect when a child is not telling the truth. This research suggests that even trained professionals, such as social workers and judges, are unable to reliably detect a child’s inaccurate statements. Some studies show adults performing at only a chance level, where others show a truth bias: adults are more likely to view children’s untrue statements as true. Having reviewed these phenomena, Part IV of the article demonstrates the ways in which the child welfare system may create, rely on, and compound, these inaccuracies. As a result, child welfare proceedings, even more than other types of cases that rely on child statements, may result in determinations based on inaccurate information. Finally, Part V of this article proposes reforms to both investigations and rules of evidence in child welfare proceedings. These proposals are two-fold: they involve changing investigation procedures to collect more reliable statements, as well as limiting the weight accorded to—and admissibility of—child statements. Ultimately, this article aims to demonstrate that caring deeply about children and their safety should not mean that the child welfare system should necessarily rely on what children say.
This section briefly outlines how cases are processed through the child welfare system, devoting attention to three notable features. First, this section addresses the sheer scope of this investigatory apparatus: child protection services captures information about millions of children every year. Second, statistics reveal a problematic level of arbitrariness when it comes to determining which families actually get caught up in the net. For example, disparity in rates of substantiation between caseworkers, a system-wide racial disparity, and a high rate of reversal show that factors other than risk to children impact child welfare determinations. Third, weak due process protections during an investigation and court case mean that families have few opportunities to correct mistakes in child maltreatment determinations. This section outlines the mechanics of investigations in child welfare cases, and explains how these investigations can lead to the termination of parental rights. The following section raises questions about the reliability of child testimony and, accordingly, the accuracy of this process.
Under federal law, every state is required to have a system in place for receiving reports of suspected child abuse. A staggering number of potential child maltreatment cases are reported to these hotlines every year. In 2011, an estimated 3.4 million reports were made to state-level child protection agencies, at a rate of 45.8 referrals per 1,000 children in the population. These referrals involved an estimated 6.2 million children.
Calls may be placed anonymously, but most reports come from professionals who are mandated to report any suspicion of child maltreatment. The majority of these reports are accepted by state hotlines for further investigation. Based on data from forty-five states, on average 60.8% of reports were screened-in (and sent for investigation) and 39.2% were screened-out. In 2011, more than two million reports of child maltreatment were accepted by state hotlines, each triggered a child protection investigation, and each received a “disposition” from the child welfare system.
Investigations are conducted by child welfare personnel employed by the state or local department of children’s services. In rare instances, the investigations also involve the assistance of local law enforcement. Investigations can be intrusive: in addition to interviewing the children in the home, caseworkers may search the home, and may physically examine the children. It is customary for caseworkers to interview children as part of their investigation, and many state statutes require caseworkers to make contact with the child who is subject of the report. A parent’s right to an attorney is not
triggered until a case is filed in court, and only a few states require that parents receive any notice or explanation of their rights. Generally, subjects of maltreatment reports consent to investigations.
At the end of the investigation a disposition is reached and the initial report is either substantiated or unfounded. In addition, regardless of whether the report is substantiated, the family may be referred to “services” to address problems identified by the child welfare investigation.
Child welfare investigations arise out of allegations of child abuse as well as child neglect. The 2011 Child Maltreatment Report determined that vastly more reports for neglect were substantiated than for abuse. The study determined that 78.5% of substantiated reports involved neglect, 17.6% involved physical abuse, and less than 9.1% involved sexual abuse. Therefore, although most of the cases that reach the national news media involve serious abuse, most child maltreatment allegations do not involve abuse at all.
Allegations of neglect may include an array of parenting infractions, including a dirty home, marijuana use outside the presence of a child, excessive tardiness and missed school days. When allegations involve the condition of the home or school lateness, the evidence needed will often involve observations about routine family business. Investigating a neglect case may require caseworkers to obtain evidence about ordinary family activities asking children questions such as: when was the last time your mother did laundry? What time does your mother wake you up for school? Or, have you ever seen your mother smoke anything? What did it look like? These statements become part of a case record maintained by the state child welfare officials. As discussed below, most research has been devoted to understanding how best to discuss serious abuse and sexual abuse with children; but the majority of child maltreatment interviews likely involve less painful topics.
After the investigation, the majority of reports to these child maltreatment hotlines are determined to be unfounded. Indeed, only one-fifth of reports result in determinations of child maltreatment. The fact that so few reports are substantiated is even more surprising in light of the generally low legal standard required for substantiation by most states. A little over half of states impose a preponderance of the evidence standard, and several other states consider a report substantiated if the investigation reveals “some credible evidence” or if abuse or neglect is to be “reasonably suspected.” That threshold does not even require a determination that it was more likely than not that the maltreatment occurred.
Although overall substantiation rates are low, the rates of substantiation vary dramatically between individual workers and between different states. At the state level, in 2011, some states substantiated fewer than ten percent of all calls, including: Virginia, Pennsylvania, Kansas, Missouri, New Hampshire, and Montana. Other states substantiated more than half: Maryland, Hawaii, Utah, Puerto Rico, Rhode Island and Georgia. Although this can be partially explained by the difference in evidentiary burden, that is not the whole story. Maryland, for example, had the highest rate of substantiation in 2011, but state law in Maryland requires the highest burden of proof: a preponderance of the evidence.
Variation also exists between caseworkers. Research on rates of substantiation has revealed that caseworker determinations may depend on a variety of factors, such as experience, self-reported skills of the caseworker, and whether the worker has a supportive relationship with coworkers. For example, a study of Child Protection Services (CPS) workers in North Carolina found the individual rate of substantiation for sexual abuse ranged from forty-five percent to sixty-three percent of cases. In another study that relied on nationally available data about child wellbeing, caseworkers were less likely to substantiate reports involving boys, even when the study controlled for other factors, such as the level of harm alleged. These findings and others have led researchers to conclude that a substantiation decision “is an imperfect measure of child maltreatment.”
Furthermore, it is important to consider significant racial disparities in the substantiation of child maltreatment reports. According to Law Professor Dorothy Roberts, “[b]lack families are the most likely of any group to be disrupted by child protection authorities.” Statistics from the 2011 Child Maltreatment study corroborate this claim. Black children were the subject of substantiated reports at a level of 14.2 children per 1000; the rate of substantiation for American-Indian/Alaskan Native children was 12.4 per 1000. White and Hispanic children, on the other hand, were the subject of a substantiated report at a rate of 8.0 and 8.4 of 1000 children, respectively. Asians (excluding Pacific Islanders) had a substantiation rate of only 1.7 per 1000 children.
Although child protection dispositions are purely administrative, the consequence of a substantiated hotline report can be significant for families. Even without appearing in a family or dependency court, a substantiated report results in the inclusion of the named adult in a statewide database. New York’s database alone contains more than 4.3 million names—meaning that a number equivalent to over half the entire population of New York City is listed in the database. A substantiated report of child maltreatment can prevent a person from obtaining a variety of jobs and can prevent them from becoming a foster or
adoptive parent. In many states, there is no state law procedure for appealing inclusion in the registry of child maltreatment. Available data suggests that seventy-five percent of substantiated reports that are appealed are expunged.
From the investigation stage, very few cases result in court action. Of substantiated reports, one study found that only nineteen percent resulted in court actions. Among the fraction of cases that are filed in court, most do not result in a foster care placement. In other words, more than eighty percent of cases where allegations are substantiated end with the inclusion of the adult in a state registry, but no formal court action. Of the cases that do go to court, most children remain with their families under supervision.
Families who do have their day in court will find that child welfare proceedings do not offer the same due process protections as have been deemed constitutionally required in criminal cases. For instance, the burden of proof in child welfare proceedings is a mere preponderance of the evidence; parents do not have a due process right to appointed counsel; and there is no guarantee of a trial by jury or a speedy trial.
Throughout the stages of a child welfare case, evidence rules tend to favor the inclusion of evidence that, in other contexts, is deemed unreliable. During many points, hearsay may be admissible. For example, when a child welfare case is initially arraigned, the court is often called on to make a preliminary determination about whether to release the child to the parent or to place the child in the temporary care of the state. These determinations can be made without a formal hearing. If the decision is made to place the child in the care of the state, the parent (and sometimes the child) may request a hearing to have the child returned. At many of these hearings, hearsay is admissible without any requirement of corroboration. In these instances when hearsay is permitted, caseworkers can testify about conversations they had with children out of court. These interim decisions can have a lasting impact on the case, particularly because months or years can go by before a trial is held on the charges of abuse or neglect.
Even when a trial is held, the outcome of a case can turn on determinations of a child’s credibility. At trial the underlying allegation of abuse or neglect must be proven by the state, but the rules of evidence are not necessarily the same as in the criminal or civil context because family courts are primarily concerned with identifying child maltreatment. A child’s hearsay statements are typically admissible only if corroborated by some other fact. Rarely, some states permit a child’s uncorroborated hearsay statement to support a finding of abuse. Yet even where hearsay is admissible with corroboration, the threshold for corroboration can be quite low. Accordingly, the evidence in a child welfare case may include a caseworker’s testimony about a prior conversation with a child. Once a finding of neglect or abuse has been entered, hearsay can be freely admissible again during the dispositional hearing.
There is no question that the child welfare system casts a wide net. The purpose of rules that welcome evidence from virtually any source and require investigation of even minor, anonymous allegations, is, ostensibly, to leave no stone unturned in the investigation of child maltreatment. One could persuasively argue that this vastly over-inclusive system is worthwhile if more cases of child maltreatment are identified. But that supposition is only valid to the extent the system is making high quality judgments. It would be problematic if the system were to intrude on private family life based on erroneous or unsubstantiated judgments. The variation in substantiation rates by state, by caseworker, and by race and gender of the child, suggest that risk factors alone do not explain the outcome of child welfare investigations. Therefore, it is critically important to look at whether determinations of risk and harm are based on the best available information.
III.EMPIRICAL RESEARCH ON CHILDREN’S STATEMENTS
Given the scope of the child welfare apparatus, the frequency that the apparatus makes contact with children, and the consequences for families, it is surprising how little data is kept about how those investigations are conducted. The empirical data that is available may shed light on some factors that impact the reliability of children’s statements. There is no question that children are
capable of accurately remembering and recounting both innocuous and traumatizing past experiences. Nevertheless, there are many reasons why a child’s account of her experiences may not be accurate.
Some of these reasons are intuitive: for example, a child’s memory is limited by her age, and any delay between the event at-issue and the relevant interview. Younger children are less likely to recall information correctly than older children, and younger children forget events more rapidly than older children. Also, a child’s age often determines whether a child understands the full import of what she is saying. For example, a four-year-old who says something happened “yesterday” may not have developed the ability to identify chronological time.
Some limitations on the accuracy of child memory are less intuitive. For example, empirical research over the past two decades has consistently demonstrated that adults may inadvertently ask questions that negatively impact a child’s ability to accurately recall her experiences. This process of unintentionally altering a child’s memory is referred to as “suggestibility.” Further, new research has begun to show that children, even young children, can be “successful” lie-tellers; children begin to tell lies during their preschool years and lie-telling behavior increases with age. The following section will focus on some of these less intuitive sources of inaccuracy.
Even ordinary conversations between children and adults can lead to misunderstandings. This difficulty that adults have understanding what children have experienced has yielded an enormous body of psychological research. Perhaps motivated by high profile cases in which children falsely accused adults of molestation, research in this area has tended to focus on childhood sexual abuse. After decades of study, research demonstrates unequivocally that suggestive interviewing causes children to make false allegations including allegations of sexual abuse. Accordingly, an unsuspecting interviewer can unintentionally cause children to make statements that are untrue. Citing this science, the Supreme Court has noted that:
The problem of unreliable, induced, and even imagined child testimony means there is a ‘special risk of wrongful execution’ in some child rape cases. . . . Studies conclude that children are highly susceptible to suggestive questioning techniques like repetition, guided imagery, and selective reinforcement. . . . Similar criticisms pertain to other cases involving child witnesses; but child rape cases present heightened concerns because the central narrative and account of the crime often comes from the child herself.
Without reviewing the entire literature here, this article aims to present a few examples of the type of innocuous questioning that can yield profoundly distorted results. Although it has been acknowledged that suggestibility is problematic, courts have generally failed to grant relief unless it is demonstrated that such suggestibility rose to the level of coercion. This judicial approach minimizes the danger posed by these seemingly innocuous questions because they can lead to false allegations and, indeed, false memories. Problematic interviews may involve interviewer bias, repeated questioning, question wording, and other subtle interview techniques.
Interviewer bias occurs when an interviewer holds a prior belief about the occurrence of some event. Studies have shown that an interviewer’s view of events can influence the type of statement they elicit. It would seem that this would not be a primary concern in the child welfare context, where an interviewer’s goal is to find out what happened to a child. That is, there is no obvious bias that would infect an interviewer’s perspective. However, as it turns out, even a slightly misinformed interviewer can influence a child’s response.
For example, in one study, preschoolers played a game similar to “Simon Says,” and one month later they were interviewed by a trained social worker “who was given a one-page report containing” both “accurate information and erroneous information” about the game. The interviewer was told that various actions might have occurred during the game and was not informed that some of the information in the report was inaccurate. Those one-page reports “powerfully influenced the dynamics of the interview.” When asked about the events that were accurately portrayed in the reports, the children correctly recalled ninety-three percent of all events. However, when asked to describe events about which the social worker had been misinformed, thirty-four percent of the three- to four-year-olds and eighteen percent of the five- to six-year olds corroborated the social worker’s misperception: that is, recalled one or more events that did not occur. Furthermore, the study reported that “children seemed to become more credible as their interviews unfolded.” Although the children were reluctant to offer details of the false events, later some children “abandoned their contradictions and hesitancy and endorsed the interviewers’ erroneous hypothesis.” Based on this and other research, experts in the field have concluded that interviewers’ beliefs about an event can influence the accuracy of children’s answers.
Researchers speculate that interviewer bias has such a profound effect on young children because they are cooperative. When questioned by a respected adult, “children sometimes attempt to make their answers consistent with what they see as the intent of the questioner rather that consistent with their knowledge of the event.” Accordingly, even a trained social worker investigating a report of child abuse may inadvertently cause a child to corroborate an event that did not occur.
The need to repeat questions when interviewing children is obvious. The interviewer may be looking for more information or may not have understood the child’s answer. The interviewer may be looking to check the consistency of the child’s report. And asking a question repeatedly can produce positive effects—when asked free recall questions (“tell me everything you remember about . . .”) both children and adults remember new information with additional interviews. But, research has shown that repeatedly asking the same question can lead to an inaccurate statement by a child.
When asked the same question more than once, “[c]hildren will often cooperate by guessing, but after several repetitions, their uncertainty is no longer apparent.” Therefore, if one asks a child a question such as whether her mother smokes, the child may answer no. But if one asks the exact same question again, just to be sure, the child is likely to believe the first answer was wrong and cooperate with the interviewer by changing the answer. True or not, the child is unlikely to revert back to the first answer, perhaps out of fear of seeming untrustworthy. With repeated questions, studies show, children omit phrases such as “it might have been” and therefore sound confident about their statements.
In addition, children are more likely to assent to a false event in a third interview than in a second interview. In fact, studies show that a child’s first interview with a neutral interviewer is the most accurate. When children add details in subsequent interviews about the same event, “these [details] have a high probability of being inaccurate.” Therefore, an interviewer can change the content of a young child’s answer merely by asking a child the same question twice. And, the more interviews that take place, the more likely a child will agree with a false statement.
There are a number of ways a question can be phrased, and studies have shown that minor differences in the phrasing of a question can dramatically impact the answer a child will give. One type of phrasing is the “forced-choice” question. A forced choice question limits the number of options (“was it red or blue?”). In one study, adults, and four-, six- and eight-year-old children witnessed an ambiguous event. The children were subsequently interviewed about what happened. The study found that the participants, particularly the four-year-olds, were likely to change their responses both within and across sessions when asked a yes/no question about which they had no information. Children, being cooperative, want to answer. Therefore, even if they don’t know the answer, children will answer a yes/no question with sheer speculation. And, as noted above, the more times they were asked the more confident they appeared in their answers, even though those answers were not the truth.
Another problematic phrasing is a “tag” question—a question that implies that the questioner knows the answer (“Amy touched your bottom, didn’t she?”). One study found that children falsely assented to more than twice as many forensic questions when they were tag questions. In the study, preschool children participated in two games, a game of Twister, and another game the researchers made up called Shapes. Half of the children were touched on the hand, arm, calf and foot, and half were never touched. A week later, children were interviewed about whether they were touched. Half of the children were asked direct questions and half were asked “tag” questions. Children who were asked a single tag question were more likely to falsely assent to the question, including questions about whether they were touched. The researchers asked tag questions about innocuous touch such as, “Did Amy touch your arm?” Or, in tag form, “Amy touched your arm, didn’t she?” They also asked “abuse touch” questions such as, “Did Amy kiss you?” Or, in tag form, “Amy kissed you didn’t she?” Finally, researchers asked more general questions, like “Did Amy give you candy?” Children falsely assented to more questions about abuse touch (49%) than non-abuse questions (20%). Therefore, the researchers concluded that children’s accuracy when answering a tag questions relevant to a child abuse investigation was at near chance levels, compared with an accuracy rate of nearly 100% among children who were asked direct questions.
Interviewers can also communicate subtle verbal and nonverbal cues in their questioning. According to the research, children are attuned to these emotional cues and will act accordingly. For example, researchers have found that even well-meaning efforts to put a child at ease can have unintended consequences. “[I]f a child is told, ‘you are a really good boy’ after he makes certain kinds of responses, the child will tend to increase these types of responses and to decrease other responses that are not reinforced.” Therefore, in interviews, “[w]hen interviewers are overly supportive of children, the children tend to produce many inaccurate as well as many accurate details.”
In one study, children were interviewed about the details of a visit to a laboratory four years earlier. In the follow-up interview, the researchers told the children they were to be questioned about an important event and asked, “are you afraid to tell?” Then the researchers said, “[y]ou’ll feel better once you’ve told.” Although most children did not remember the original event, in response to the interviewer’s subtle cue that something had happened, some children “falsely reported that they had been hugged or kissed, or that they had their picture taken in the bathroom, or that they had been given a bath.”
The problem for interviews in the child welfare context is apparent: where an adult is interviewing a child suspected to be the victim of abuse or neglect, there will be strong emotional pull to support and encourage the child. This most natural reaction is precisely the type of subtle cue that can lead to inaccuracy. In fact, even if a child receives only one small inaccurate suggestion, it can lead that same child to make to greater fabrications. The more times a child is interviewed the greater the likelihood that the interview will result in a statement that does not reflect the truth about that child’s experience. Additionally, the research on suggestibility has tended to show that small changes in question form or atmosphere can alter a young child’s statement.
For older children, the problem of suggestibility is different but not absent. Older children may be better-equipped to resist cues—in part because their understanding of memory and time is more developed—but they are still more susceptible to coercion than adults would be, and this can be considered a form of suggestibility. Certainly, scholars have noted the problem of extreme coercion that can occur in a custodial police interrogation of a minor. But in the non-custodial setting of a child welfare investigation, inaccuracy among older children may take the form of lying.
Researchers have demonstrated that children lie regularly. Indeed, research shows that children begin to tell lies during their preschool years and lie-telling behavior increases with age. By age three, children are already “successful” lie-tellers. Although lying is considered a bad behavior, it is also a natural part of growing up—indeed, studies have shown that lie-telling behavior is “associated positively with children’s cognitive development in terms of their understanding of other’s minds.”
We do not know why children lie. In a real world context, lying can be difficult to measure. In the laboratory setting, children’s lie-telling behavior is most often tested using a temptation resistance paradigm: children are exposed to a temptation like peeking at a toy in a box or answers to a test, are instructed not to peek, and then are later asked if they did peek. The temptation resistance studies measure lying to conceal a transgression and reveal that most children lie about transgressions.
In one study involving 123 children between the ages of 3 and 8, 82% of children peeked at the toy in the box, against instruction, and 64% of those who peeked lied about peeking. However, most of the lying children within this age group were not able to maintain their lies: of the seventy-nine children who said they didn’t peek, seventy-two percent later responded correctly when asked to identify the toy in the box—a fact they only could have known if they had peeked. When asked about this discrepancy, forty-nine percent of the children provided plausible explanations that allowed them to maintain their lie. Older children were more successful at maintaining the lie: among elementary school students, seventy-seven percent of the lying group gave plausible explanations to maintain their lies. This ability to tell lies to cover other lies emerges around seven to eight years of age.
Older children lie about their transgressions at a higher rate. In a study of adolescents, sixty-eight percent of children peeked at the answers, against instruction, and eighty-two percent of those who peeked lied about peeking. The study was repeated, with one change: students were asked to promise to tell the truth. Following the promise, sixty-five percent of the peekers continued to lie. Thus, when given an instruction to be truthful, children still lie at high rates. Other research suggests that children are especially likely to lie when they would not be implicated in the transgression. This is particularly concerning for the child welfare context where children may be assured that nothing bad will happen to them. Furthermore, empirical studies have demonstrated that children can be coached by adults to lie. The research also reveals that children can be coached to lie, persuasively, about emotionally significant events.
One study regarding adolescent lying is particularly illustrative. The study required adolescents to self-report lies they told to their parents. Researchers found that adolescents frequently lied about drugs/alcohol, parties, friends and money. Eighty-two percent of participants indicated that they lied to their parents about at least one issue in the past year. These numbers likely underestimate the extent of lie-telling because of the negative perception associated with that behavior. The authors reasoned that lying might be more prevalent during adolescence, as children seek to gain increased autonomy within the family. Indeed, the study found that sixty percent of participants believed that “assertion of an autonomous right to make decisions” was an acceptable motive for lying. Finally, a child may intentionally lie because they are coached by others. Interestingly, one study of preschool and elementary school children found that elementary school children were more susceptible to coaching than preschoolers.
Therefore, to the extent the child welfare system assumes that children are unlikely to lie at all, that assumption is not supported by the research. As discussed above, the research suggests that children may be motivated to lie in situations where the lie would help the child avoid punishment or achieve autonomy. Although there has been little research examining whether children may spontaneously lie about something as serious as child abuse, it is important to note that a child’s lie need not involve something as significant as sexual abuse to have a major impact. Where neglect cases can turn on facts such as whether a child was ever left alone, or whether a child was disciplined using a belt or the hand, a child’s exaggeration may be enough to support a finding of child maltreatment.
In the previous sections, this article focused on suggestion and lying as two reasons why a child might give an inaccurate statement to an adult. However, the risks of inaccuracy can be mitigated if adults in the child welfare system can readily detect discrepancies, or are willing to accord less weight to child testimony in light of these risks. This section suggests that, in practice, the opposite is true: research demonstrates that adults cannot easily detect children’s lies and tend to view children’s statements with a “truth bias.” It has often been said that “it is easy to detect false reports that are the result of suggestion, because . . . children . . . merely ‘parrot’ the words of their interrogators.” However, empirical research provides no support for this assertion. On the contrary, trained professionals in the fields of child development, mental health, and forensics tend to view child reports as highly credible, even when produced by suggestive interviews.
One study asked adults to evaluate child reports when children were asked (non-suggestive) free-recall questions and (suggestive) yes-no questions. Before watching the interviews, some of the adults listened to a presentation about suggestibility that was originally created to train child protection workers on proper investigative interview methods. The study found that all participants, regardless of training, “believed the majority of children’s event narratives regardless of whether those narratives were volunteered in response to open-ended questions/invitations or elicited after ‘yes’ responses to yes-no questions.” The most common reason participants offered for believing an event narrative “was the simple fact that there was a narrative.” Even though the participants knew they were in a psychology experiment and that the children had been exposed to misinformation, participants thought that false “yes” answers were accurate eighty-four percent of the time, compared to an actual accuracy rate of only sixty-two percent. Therefore, according to the authors, “the present studies help us understand why cases with major inconsistencies and improbable events in children’s testimonies sometimes progress through the child protection and legal systems: because adults usually believe young children.”
Other studies show adults cannot detect lies in children beyond the level of chance. In one study, adult participants watched videotapes of children testifying about an event in a court simulation. Prior to the court simulation, parents were instructed to coach their children to tell a story about an event that did not occur, and then children were instructed to tell a true or false story. The children “were instructed to be as convincing as possible” and try to make others believe their story was true. In the simulation, an individual posing as a prosecutor prompted the children with open-ended questions, and then the children were cross-examined by an individual posing as a defense attorney. The study found that “adults were only able to detect children’s coached true and false reports at chance level.” Furthermore, adults tended to approach the situation with a truth bias: before cross-examination, adults were more accurate at detecting truth-tellers (74%) than lie-tellers (25.8%), but their overall accuracy was at chance level (49.7%). The researchers determined that “in general adults are unable to discriminate between true and false reports in children merely by observing the children testify.” The ability of adults to detect lies remained at chance level after they saw children give fairly lengthy accounts of past events and answered follow-up questions—that is to say, observing cross-examination decreased the “truth bias” but failed to increase detection of truth beyond the level of chance.
Another study examined whether adults’ lie-detection ability was improved when the adults were presented with reports of children who had actually been involved in a stressful, emotionally salient event, as well as reports of children who pretended to be affected by an emotionally salient event. The researchers found that “lay judges were unable to distinguish the reports as truthful rather than fabricated.” Of all the reports, laypersons were most inclined to believe those that had been fabricated through coaching by the child’s parent(s). In sum: participants only correctly judged children’s reports at the level of chance, but tended to believe the children who were coached at a higher rate than children who were telling the truth.
However, some studies have also looked specifically at the ability of judges or other trained professionals to evaluate the veracity of child testimony. These studies are particularly important in the child welfare context, where fact-finding determinations are almost always made by judges and not juries. These studies illustrate that experience interviewing children does not necessarily improve adults’ ability to detect truth from lies.
In one study that measured the lie-detection ability of Swedish judges, police detectives and laypersons, the only group that showed an ability to discriminate correct from incorrect witness responses was the police detectives. However, even at their best, the detectives made a substantial number of incorrect judgments and some of them performed at chance levels. Furthermore, all three participant groups had a strong tendency to judge a statement as correct rather than incorrect, and this truth-bias was particularly pronounced among the judges. Similarly, another study compared the lie-detecting ability of trained experts with individuals that did not have any training, and found that experts demonstrated a stronger overall predisposition to believe children. The study also showed that experts had confidence in their (as it turned out, inaccurate) judgments.
Some studies have focused specifically on the ability to detect lies in hearsay testimony. While hearsay testimony is generally prohibited in criminal and civil cases (with exceptions), hearsay is admissible in child protection proceedings. In a study of mock jurors in child sex abuse cases, jurors rated hearsay witnesses as “more confident, accurate, and consistent and less suggestible than the child witness.” Researchers have found that mock jurors are impacted by hearsay testimony to a considerable degree.
Research has also shown that the ability to judge accuracy is generally better when testimony is presented in written rather than in visual format. One possible explanation is that “something in the visual modality interferes with the detection of accuracy cues.” For example, visual cues (e.g. body language, eye contact) may misguide judgments of statements presented in the videotaped display. Accordingly some have concluded that, “judgments of accuracy based on transcripts rather than live testimony would increase the quality of legal decisions.”
The emotional intensity of child welfare proceedings may exacerbate the effect of truth bias on adults’ perception of child testimony. Researchers have evaluated the impact of emotional intensity on adults’ ability to detect untrue statements in other adults. In one study involving adult accounts of sexual assault, observers became less accurate in their assessments of witness testimony as the level of emotional intensity of an account increased. The researchers concluded that “a common pattern among evaluators is one of mediocre performance combined with inflated confidence, a formula for disaster in forensic settings.” Because of the emotional intensity of child welfare proceedings, there may exist a similar bias towards assuming the truth of the alleged victim testimony in these emotionally-charged cases.
In sum, what is known about children’s testimony strongly suggests that an adult’s judgment about a child’s truthfulness is not to be trusted. When asked to evaluate the truthfulness of child testimony, research suggests that caseworkers are wrong in a significant number of cases: wrongly discrediting some of the children who were abused and wrongly crediting some children who were not. The risk of such inaccuracy alone should prompt an interrogation of the current system.
In a child welfare investigation, the focus is primarily on the wellbeing of the child. It follows that children’s statements are perhaps more likely to be made a part of a child welfare case than other types of cases. Although the fundamental right to parent is at stake, child welfare investigations may ultimately rely on something a child has said without a process for testing the accuracy of that statement.
This section will address some unique features of child welfare proceedings that discourage the rigorous testing of children’s statements to ensure their accuracy. As a result, important child welfare decisions may be made based on evidence of questionable veracity. This section will examine the inadequate attention to interview technique, the inclusion of hearsay testimony from children at most stages of the litigation, the role of attorneys for children, the pressures on judges deciding child welfare cases, the lack of meaningful truth induction opportunities, and the difficulties created by the need to offer children therapy or other interview-based mental health services. The final part of this section includes a case that exemplifies the consequences of this flawed system.
One of the core duties of a child protection worker is to interview children. Proper interview practice should be a concern for all parties involved in child welfare practice. Although suggestive interviewing has led to some high profile false-positives, the most common suggestive techniques also suppress details from children who have in fact been abused. Despite the importance of getting these interviews correct, by and large, interviews are not being conducted properly.
To begin with, there is a “right” way to interview children to minimize suggestibility problems and maximize accuracy. There is “surprisingly broad international consensus regarding optimal interview practices.” For example, as one researcher has described, there is a “universal emphasis on the value of narrative responses elicited using open-ended prompts.” Open-ended prompts may include questions such as, “Tell me everything you remember about your holiday vacation.” Those types of questions do not introduce new information and avoid the pitfalls of forced-choice and tag questions. Various jurisdictions and agencies have developed structured interview protocols to guide interviewers through the questioning of a child. The goal of such protocols is to avoid problems of suggestibility. In addition to guiding the interviewer to ask open-ended questions, these protocols typically include specific instructions for rapport building to avoid the need to improperly encourage children. Also, these protocols typically involve setting up ground-rules to help a child feel comfortable saying “I don’t know.”
Despite overwhelming scientific support for interview protocols, only one state (Michigan) requires child protection workers to use a structured interview protocol and no jurisdiction has linked the use of a protocol to admissibility. Some states rely on training caseworkers. The majority of states make no additional effort to conduct interviews using empirically tested methods. Four of the five states that had the highest number of child welfare cases—California, Florida, Texas, and New York—do not require caseworkers to use any interview protocol. Together these four states accounted for nearly 1.5 million dispositions in 2011.
Instead, research demonstrates that, in the child welfare context, best practices are frequently not employed. One study looked at forensic interviewers who were trained in actual child abuse investigations and found that, despite their training, forensic interviewers formatted their questions in a problematic way about sixty-six percent of the time. Additionally, despite their training, interviewers abandoned other best practices such as establishing ground rules, building rapport, and practicing a narrative response. Another study, in England, asked specially trained police officers to interview children who had witnessed an event in their school one month prior to the interview. Approximately ninety percent of the interviewers reported that they had received training in investigative interviews and interviews of children. The study found that the interviewers asked the most suggestive types of questions approximately twenty percent of the time. Within this study, approximately one-third of the leading questions introduced inaccurate information. Even when interviewers are warned to not ask leading questions, research suggests that suggestive questioning may occur. For example, one study found that, though the interviewers were instructed to avoid leading questions, thirty percent of all interview questions could be characterized as leading, and half of those were misleading. Most problematically, “children agreed with forty-one percent of the misleading questions.”
The risk of suggestive interviewing is compounded by the generally poor quality of training provided to caseworkers, which has received media attention. This is not necessarily intentional or negligent—rather, research shows that it is incredibly difficult to train interviewers to ask non-leading questions. Even absent high caseworker turnover and limited training resources, training on proper interview technique is very difficult to impart. According to one researcher, “[e]fforts to train experienced interviewers to reduce the number of leading questions and yes-no questions asked have been largely unsuccessful.” Two studies found that after a ten-day professional training program, nearly seventy percent of the questions asked during the interviews were yes-no questions and thirty-four percent of questions could be characterized as leading. According to a group of researchers in this area, “even the most intensive training programs . . . have little, if any, effect on the actual behavior of forensic investigators.” The only forensic protocol that has shown to reliably generate open-ended questions is the structured interview protocol developed by the National Institute of Child Health and Human Development (NICHD). The structured protocol reduces much of the need for interviewers to formulate their own questions.
The problem of mishandled interviews creates acute difficulties for legal professionals. For example, once an improper interview has been conducted, it is difficult for attorneys to demonstrate the benefit of proper interview methods because no one can know what a proper interview would have produced. Similarly, it is difficult for attorneys to prove that suggestive techniques have, in fact, been used. Further, an interviewer’s in-court testimony may tend to minimize the amount of suggestive technique used. Studies show that even when interviewers have their notes available, they tend to have difficulty reconstructing their interviews with children. Furthermore, as discussed above, interviewers ask leading questions even when they are trained not to do so, and interviewer may not be aware that he employed a suggestive technique. Other researchers in this area have concluded that, “interviewers are biased to recollect that they ask nonleading questions,” and provide “inflated estimates of the open-ended as opposed to closed questions they ask.”
The inability of interviewers to accurately reconstruct child interviews has been empirically demonstrated. In one study, twenty-seven experienced professionals interviewed preschool children about an event all the children had experienced. Immediately after, the interviewers were asked to recall their interviews and their recollections were compared to actual interview transcripts. On average, interviewers correctly recalled eighty-three percent of the major event activities, sixty-five percent of the event details, and sixty percent of the errors that children had reported to them. However, the interviewers only recalled twenty percent of specific interview questions and answers. This suggests that interviewers are unlikely to reveal whether they used leading questions or employed other techniques that could have produced misinformation. Accordingly, the authors of the study posited that hearsay testimony provided by interviewers of child witnesses is “highly likely to contain the gist of the information obtained during the interviews, less likely to contain details, and highly unlikely to retain specific, verbatim questions and answers.” Furthermore all of these factors make it difficult to prove, after the fact, that an interview was flawed.
In many jurisdictions, even children’s out of court statements can be considered for the truth at child protection hearings. Hearsay statements are generally precluded from both civil and criminal trials, because they are “presumptively unreliable.” There are exceptions to the rule against hearsay, which allow for hearsay statements to be admitted if they possess certain indicia of reliability. In criminal cases, testimonial hearsay evidence that falls within such an exception must additionally comply with the Sixth Amendment’s Confrontation Clause, which accords a criminal defendant the right to confront witnesses against her.
Despite these concerns, many jurisdictions allow for the introduction of a child’s hearsay statements in child protection hearings. In some states, the uncorroborated hearsay statement of a child, standing alone, is sufficient to support a finding of abuse. Therefore, in the child welfare context, an unintentionally suggestive interview with a child may produce an inaccurate response. But, through hearsay, the child’s statements can be introduced without facing cross-examination. And, as discussed above, the person offering the child’s statement is unlikely to accurately recall potentially suggestive interview questions that produced the statement. Therefore, admitting hearsay decreases opportunities in the system to correct misinformation. And, research shows that the hearsay testimony may be have a powerful effect on the listener. As discussed above, research shows that adults tend to find the hearsay statements of children to be reliable. Accordingly, the introduction of hearsay statements in child welfare cases raises grave questions of fairness.
In some states, children are entitled to lawyers in child protection proceedings. In other states, courts have discretion to appoint counsel for children. It may be suggested that problems of reliability and suggestibility may be mitigated if attorneys are appointed for all children in child protection proceedings. However, there are reasons to believe that this may not be the case.
As noted by Professor Annette Appell, “attorneys are unlikely to share the same socio-economic background, cultural values or kin as the children they represent.” Appell notes that this professional relationship creates a confidential space between attorneys and children. “In this space, the attorney has relatively free reign to identify and shape the child’s interests and little accountability when acting within this attorney-client-child relationship.” Empirical research supports Appell’s concern. Because of our interest in protecting children, and the potentially sensitive issues implicated by child welfare cases, lawyers may be inclined to raise painful topics gingerly. This may lead lawyers to create the type of “overly-supportive” environment that “tend[s] to produce many inaccurate as well as many accurate details.” Despite these risks, a widely attended national conference on child representation that drafted recommendations regarding communication with child clients did not address concerns around suggestibility at all. The conference did recommend that attorneys for children “utilize verbal and nonverbal methods to communicate that it is safe for the client to discuss sensitive or private matters.” Yet, as discussed above, these nonverbal cues might actually exacerbate the suggestibility of interview questions. The very same instincts that cause attorneys for children to care for and want to protect their clients can create reliability problems.
Lawyers for children, sometimes solo practitioners, are not likely to receive any of the training that even caseworkers receive when it comes to proper interview technique. Further, the structured interview protocols are geared towards caseworkers, and not attorneys, and there are no known attorney protocols for reliable interviewing. After an attorney’s interview is complete, opposing counsel would have no way to measure the impact of that lawyer’s interview on the child’s statement. Content of a lawyer’s interview with a child would be privileged.
Finally, attorneys for children have a role in preparing their clients to testify in court. As discussed above, empirical research has shown that adults believed three out of four coached reports. Even though some of these coached children may be telling the truth, merely prepping a child witness may interfere with a fact-finder’s ability to evaluate that child’s truthfulness.
Judges in child welfare cases are often required to make determinations under fraught circumstances. The costs of failing to identify a case of child maltreatment are enormous, both to a child who is injured and to the deciding judge’s reputation. Professor Martin Guggenheim notes that child welfare decisions are rarely criticized in public for wrongfully ordering the removal of a child. Instead, the media focuses its attention on those rare, but extremely sad, “false negative” cases, where the state does not intervene in time and children suffer serious harm, or death. Guggenheim notes that this attention creates pressure to “err on the side of safety,” and the prevailing culture offers emotional rewards for children’s lawyers to play a “heroic” role in rescuing children from risk. But, what impact does this pressure have on a judge’s ability to learn the test a child’s statement for the truth?
Judges in the child welfare context receive little guidance, beyond their experience and intuition, about how to reliably evaluate a child’s statement. For example, in one case, a judge was asked to determine whether a teenager was biased against her stepfather. The court found that, while the child expressed anger against her stepfather when denied computer privileges, “the force of this bias appear[ed] no greater than any teenager would have for a parent asserting limits.” Without any guidance, this judge did not appear to interrogate whether the force of this bias (even if perceived to be normal in comparison to other teenagers) might cause a child to lie. More research would be helpful to guide courts in precisely these types of situations.
As discussed in this paper, there exists a significant amount of empirical research that might help guide judges in their evaluations of child testimony. However, due to busy court calendars and the need to qualify expert witnesses, judges in most cases will likely not have the benefit of the available science. Set against the high cost of failure, judges may tend to err on the side of caution when evaluating a child’s account.
Truth induction refers to those processes the legal system employs to encourage truth-telling. In the context of a child’s in-court testimony, typically these methods include a competency hearing, an oath, and cross-examination. There are number of problems with the use of these truth-induction regimes in child welfare proceedings. As discussed above, asking a child to take an oath or promise to tell the truth is an ineffective way to encourage truthful statements. Empirical studies have demonstrated that in children ages three to eleven, the effect of “oath-taking” on truthfulness is “moderate.” Studies suggest that sixty-four percent of younger children (ages three to eight) and forty-eight percent of older children (ages eight to sixteen) continued to lie even after promising to tell the truth. Other forms of truth induction, such as reassuring children there will be no punishment for telling the truth, have been equally unsuccessful at encouraging truth-telling; as discussed above, children are more likely to lie when they feel they would not be implicated in the transgression. Accordingly, relying on oath-taking is not sufficient to meaningfully encourage truthfulness.
Likewise, a competency hearing is unlikely to ensure that a child witness will make truthful statements. The purpose of a competency hearing is to determine whether a child knows the difference between the truth and a lie. Yet one study of children’s lie-telling behavior found a very weak relationship between a child’s conceptual understanding of truth and lies on the one hand, and the child’s actual behavior on the other. That is, a child may be able to tell the difference between a truth and a lie but that ability does not prevent her from telling lies.
Recently, legal scholars have identified a trend towards lowering the bar in competency hearings. Further, when a child comes to court to testify, particularly about an emotionally significant event, the court and attorneys may be motivated to shelter the child. Therefore, a competency hearing is unlikely to be a searching inquiry into the child’s capacity to testify truthfully. In terms of any effect on the fact-finder, research suggests that viewing a competency hearing does not improve the ability to distinguish a child’s truth from lies.
Finally, cross-examination of children presents complicated challenges. In child protection cases, there is a general reluctance to cross-examine children aggressively or even at all. Children may be thought of as victims such that “[w]hen they appear in court” they may be “protected from the harshness that so often characterizes the legal process.” Additionally, cross-examination is not an effective way to discredit a child: as others have discussed more fully, it is hard to imagine a judge persuaded even after a lawyer succeeds in pinning a child to a prior inconsistent statement or questioning a child’s motive. In a proceeding designed to protect children, a lawyer’s aggressive examination is unlikely to sway the fact-finder.
Another challenge posed by cross-examination is that it may include the kinds of questions that facilitate inaccuracy, such as leading questions, complex language, and tag questions. Indeed, studies have shown that children exhibit a high degree of compliance with leading questions on cross-examination and rarely ask for additional clarification, even in the face of a complicated question. In fact, in one study, eighty-five percent of children changed at least one aspect of their original testimony during cross-examination and one third of children in that study changed all of their previous responses. Research in this area illustrates that cross-examination does not help to reveal inaccurate child statements and can solicit inaccurate testimony from children whose original reports were accurate. These results are unique to children because of factors related to youth, including: low levels of linguistic competence, susceptibility to intimidation, a tendency to assume that an adult questioner knows the answers, and lack of memory.
Many children will be engaged in mental health treatment during the pendency of any child protection proceedings. Indeed, one of the primary goals of the child protection system is to engage the family in services. Often children will receive mental health treatment “while still involved in protracted legal proceedings.”
Because the goals of a forensic interview and the goals of a therapeutic session differ so greatly, statements made in the course of therapy are a problematic source of evidence. A therapist, for instance, may have very good reasons to encourage certain responses and praise children for their disclosures. Some of these interview behaviors may encourage misinformation. The literature shows that therapists demonstrate interviewer bias in that they “rarely test alternatives, and fall prey to illusory correlations and confirmatory biases.” Where post-traumatic stress is at issue the best available therapeutic treatment will often involve talking about the traumatic event.
Given the importance of promoting a child’s mental health and emotional wellbeing, it makes little sense to delay therapy. However, in order to preserve the integrity of child testimony, professional organizations recommend that forensic interviews of children “be conducted separately from therapeutic efforts, in separate sessions by different professionals, often with limited sharing of information between the two.” Therefore, researchers in this area caution that “suggestibility findings heighten concern that if child witnesses engage in therapy while involved in lengthy legal proceedings to treat post-traumatic symptoms, therapists will use questionable techniques, creating false allegations or distorting genuine memories.” Indeed, therapy based on inaccurate information from children can reinforce a falsehood and create psychological confusion for a child that may be difficult to repair. Despite these empirically-substantiated concerns, “[t]here is virtually no research on how to meet child witness mental health needs without tainting their reports.” Accordingly, courts have a difficult time sorting out allegations that arise out of a therapeutic setting.
One case, In re Jeffrey S., illustrates several problems with child testimony that have been raised in this article. In this case, an Ohio appellate court found that two children were denied any contact with their parents for four years based on the false belief that their parents sexually abused them—a belief that was facilitated by a combination of improper interview technique and botched case handling.
The case involved two children, Jeffrey and Benjamin. The older child, Jeffrey, was born prematurely, was small for his age, had some special needs, and was diagnosed with Dwayne’s Syndrome. It is undisputed that his mother sought help for him when he was in her care. At a certain point, however, Jeffrey’s school expressed concern about his behavior, including potential symptoms of sex abuse. Jeffrey and Benjamin were removed from their mother’s care and placed in foster care. Three months later their mother consented to a finding of dependency, but not abuse or neglect.
Essentially, Jeffrey’s mother admitted that her poverty prevented her from caring for her son. Following the disposition, a case plan outlining the parents’ obligations and requirements for reunification was filed. Less than a month after being placed in foster care, Jeffrey was sexually molested by another foster child while in the custody of Children’s Services. Months later, still in foster care, Jeffrey began receiving individual, family, and group counseling to deal with the molestation which occurred in the foster home. Jeffrey was described by his therapists, doctors, and caseworker as “very creative, bright, manipulative, and highly imaginative.”
Benjamin, three years old at the time he was removed, was not given immediate counseling to deal with separation issues, despite requests by the parents. According to an assessment by counselors for Children’s Services, Benjamin did not require counseling since he appeared to be “happy and well-adjusted.” Benjamin was enrolled in a facility for handicapped children as a “peer model,” an example of model behavior. Visits between the children and their parents occurred consistently; these visits were considered by children’s services workers to be very positive and the children were eager to see their parents, especially their mother. The children engaged in crafts and games, which were provided by the parents.
Six months after beginning therapy, Jeffrey stated to his therapist that his mother had “kissed his pee pee,” but then immediately recanted, saying that it was a lie. Based upon this supposed disclosure, the Children’s Services caseworker began an investigation, which resulted in allegations that Jeffrey’s mother was telling the children not to speak with counselors. The amount of visitation between the children and their parents was curtailed and closely monitored by two trained supervisors from a local sexual abuse organization. These supervisors later testified that the visits were positive, the children were eager to see their parents, and the children did not want to leave at the end of the visit. Children’s Services transportation employees later testified that when transporting the children together they encountered no problems and that the children appeared happy to see their parents. Despite these reports, shortly after visits were curtailed, the court heeded the recommendation of the children’s guardian ad litem and entered a no contact order between the parents and children. The children’s communications and visits with both parents ceased, even though there were no allegations that the father had discouraged the children from discussing any abuse.
Both boys were placed in group therapy sessions with children who were “are all disclosing sexual abuse.” Jeffrey attended a group that required each child to identify himself at the beginning of the session, say why he is in the group (for example, “I was sexually abused”), and identify who committed the abuse. Jeffrey refused to join in the circle, threw things, and cried. When he finally spoke to the group he indicated that he was in the group because of sexual abuse by a foster brother, and did not name any other perpetrator. By the end of a year he had become a group leader and improved his social skills; he had also “disclosed” that he had seen his mother’s breasts and they were big. He also once mentioned briefly that his mother had tried to have sex with him, but gave no details. Based on these allegations of sexual abuse, Children’s Services moved for permanent custody of both boys. A hearing took place over sixteen days throughout a seven-month period. The state presented testimony from various witnesses as to the children’s therapy regarding the allegations of abuse—some of which were patently false. For example, Jeffrey supposedly disclosed that his penis had been cut off and he had been taken to the hospital where it was reattached. In another “disclosure,” Jeffrey alleged that, during a closely monitored visit at children’s services, his mother had taken him to the bathroom and “kissed his pee pee.” The Children’s Services worker admitted that this could not have happened and further acknowledged that they were unable to determine whether or not sexual abuse by the parents had ever occurred. Both boys had been examined and showed no physical signs of abuse. The children’s services caseworker testified that both parents successfully completed all the parenting programs and had attended therapy as directed. The caseworker stated that the main thing preventing reunification was the parents’ inability to “acknowledge that the boys have been sexually abused and . . . to put safeguards in their home to make those children feel safe.” Although the caseworker acknowledged that Jeffrey and Benjamin were traumatized by the upheaval of this process—Jeffrey had been in five different placements in the previous three years, and Benjamin had been in four—she testified that she could not recommend reunification until the parents would admit that they sexually abused the children.
At the hearing, Jeffrey and Benjamin’s parents presented evidence that the children’s supposed disclosures were, in fact, false memories. The mother called an expert in forensic, clinical, and family psychology in support of this argument, but the trial court granted permanent custody of both children to Children’s Services. The parents appealed. It was only after reviewing more than 3,500 pages of transcript, and numerous other documents and reports, that the appellate court was able to untangle what had transpired, and reverse the trial court’s determination. Two years after the trial court granted permanent custody to Children’s Services, four years after the children were barred from any contact with their parents, and six years after the children were put in foster care, the case was resolved.
This case presents, perhaps, the perfect storm. The initial allegation was made by a mandated reporter at Jeffrey’s school. Once those initial allegations were presented in court, the parents admitted that their poverty prevented them from providing for Jeffrey. Although no court found that the parents were unfit, the children were exposed to repeated contact with various professionals. Jeffrey made his initial allegation of abuse to a treating therapist in an interview that was not guided by forensic procedures. Jeffrey’s statements were offered at trial against his parents as hearsay. Despite hearing from an expert forensic psychologist who questioned the veracity of Jeffrey’s disclosures, the court credited Jeffrey’s comments to his therapist. Finally, the no contact order that prevented the parents from visiting their children for four years was entered at the suggestion of the child’s guardian ad litem who was unable to detect that Jeffrey’s statements were not true. Delays in the proceedings, including a sixteen-day trial spread over seven months, prolonged the separation of the parents from their children.
The facts of this case, though extreme, reflect the current legal framework of most child welfare cases. For the reasons stated above, this legal framework is ill-equipped to evaluate children’s statements. Rather, the structure of child welfare proceedings may in fact facilitate inaccurate child testimony. As a result, vital child welfare determinations may rely on children’s untruthful statements.
In light of the problems discussed above, this section aims to outline a few possible solutions. The following proposals focus specifically on improving the system’s ability to collect and filter children’s statements. Some of these proposals are derived from the criminal context, and others merely parrot the recommendations that researchers in the area of forensic psychology have been making for years. The proposals fall into two categories: those aimed at collecting better statements from children and those aimed at improving legal processes that adjudicate these allegations.
In response to concerns about interview form, many jurisdictions have created Child Advocacy Centers (CACs) where teams of professionals including medical personnel, child welfare social workers, and police can collaborate to interview a child who may be the victim of a crime. The referral procedures for CACs vary, but generally children are only interviewed at a CAC when allegations appear serious enough to warrant a possible criminal charge. These are, typically, cases involving allegations of sexual abuse and serious physical abuse. Accordingly, the vast majority of child welfare investigations do not pass through CACs. And for good reason. It is disruptive to children’s lives to be taken to a CAC, examined by a doctor and interviewed by a police officer—such a procedure would be overkill if applied to every case in which the allegations merely suggest the parents kept a dirty home. Furthermore, CACs lack the capacity to handle the enormous number of cases that are reported each year. While a full CAC assessment may not be warranted, or even possible, interviews in the child welfare context might benefit from some similar features.
One relatively straightforward recommendation is that policy makers should consider requiring structured interviews during the initial stages of a child welfare investigation. As discussed above, the structured forensic interview protocols developed by the NICHD facilitate more reliable investigations. Structured interviews can help generate truthful statements by establishing ground rules for the interview that motivate truth-telling.
Yet, there are several constraints that may prevent the use of structured interview techniques. When making contact with a family, child welfare investigators must achieve an extremely delicate balance. They must knock on the doors of complete strangers, ask to be let inside, and then begin to discuss deeply personal matters. They lack the uniform—and the side-arm—of police officers. They are aware that failing to detect abuse could have dire consequences. They must simultaneously seek cooperation from adults in the home while demanding personal and family information. They have to transition quickly between talking to adults and talking to children. They have to be comfortable talking with children of all ages, from toddlers to teenagers. Before speaking with a child, they likely do not know if that child has a disability, for instance, or if that child speaks English. When they sit down with a child they have no idea whether that child is afraid of being taken from her family, or wishes, more than anything, to be taken from her family.
Despite the complexity of this task, society provides little education, training, or compensation to child protection investigators. The goal of initial case contacts is to ensure the safety of children in the household. Still, in many cases those initial case contacts can yield significant information.
One model developed in Oregon attempts to address the need to apply forensic interview standards to field investigations. The 2012 edition of the Oregon Interviewing Guidelines defines two distinct interview types: forensic and initial responder. Forensic Interviews “are conducted in a manner that is legally sound, of a neutral, fact-finding nature, and coordinated to avoid duplicative interviewing.” An initial responder interview typically takes place during the first contact with the child involved in the child welfare investigation. The guidelines recommend that initial responders should make every effort to limit the number of times a child is interviewed and note that often enough information is available from the reporting source to avoid an initial interview with the child. Under the protocol, the initial contact should only gather as much information as necessary for the protection of the child. By limiting the goals of the initial interview, the Oregon protocol allows the majority of in-depth questioning to occur at a later forensic interview. The Oregon protocol suggests that initial interviews with children should be limited to those situations where speaking with the child is necessary to establish safety or determine the need for criminal and medical intervention but that it is sometimes possible to avoid a full investigatory interview during the initial meeting.
The Oregon protocol also contains some guidance for how to talk to children. The protocol suggests that, “if the child volunteers detailed information, that information should be written down or otherwise recorded, and the report should reflect the circumstances under which the child made the disclosures.” Further, “if the child is not volunteering information, the initial responder should avoid questioning her, particularly asking leading questions, and the information needed should be obtained from sources other than the child whenever possible.” The protocol offers other helpful guidelines such as: “Do not ask the child why the abuse happened, as it implies to her that she is to blame.” Unfortunately, the Oregon protocol itself notes that “it should not be taken as a dictate . . . that every interview in Oregon must follow this format.”
Since 1998, Michigan has required all interviews with children to follow a standardized forensic protocol. The required standardized protocol was developed by a Michigan Governor’s Task Force. Given how long Michigan has been working on this project, it is somewhat surprising that no other states have required forensic interviewing by law. Other states could adopt a protocol similar to Michigan’s, or the NICHD model, which is similar but was developed by independent researchers and has the benefit of empirical support. A number of other steps can be taken to improve interview quality. Even without formalizing a structured interview, recording all of the initial case contacts that currently take place would at least provide an opportunity for supervisors and researchers to review how these interviews are conducted. That way, should a child disclose detailed information, her exact statement would be preserved. Nearly all structured interview protocols recommend that these interviews be recorded. It would require very little extra investment since portable recording devices are widely available in the era of smartphones. Recording would allow the interviewer to timestamp the interview and to instantaneously share the tape with supervisory staff.
Alternatively, if legislators do not want to modify the child welfare system to require videotaped interviews, judges could require proof of proper interview form before admitting a child’s statement. Or, even without a statutory requirement of a recorded statement, rules of evidence could be modified to admit only forensic interviews that have been videotaped. Accordingly, the interviews could be conducted however child welfare professionals think is best, but they would be barred from admitting any unrecorded hearsay statements produced by these interviews. Such rules would remove pressure from jurists who are otherwise tasked with untangling various statements made to various people under uncertain conditions.
Whether through child welfare legislation, legislative modifications of evidence rules, or individual court requirements, tape recording of these interviews is necessary to ensure fairness and accuracy. If initial interviews are not recorded, there will be no way to uncover any inaccuracies the interview may have generated, or to challenge the use of suggestive techniques.
Accordingly, the available research, though useful, does not provide a perfect model for interviews in the child protection context. Additional research on minimizing suggestibility and maximizing accuracy during the initial meeting with a child is needed. Many existing structured interview protocols appear to be based on learning whether a particular significant event occurred in a child’s life. However, initial interviews in the child welfare context are likely to involve less obviously significant or traumatic events. Further research may help states develop protocols that are better adapted to the constraints and purposes of child welfare investigations. Nevertheless, such protocols offer a promising alternative to current methods.
As discussed above, in child protection cases, the law asks judges to test the accuracy of a child’s statement in court (or in the form of an admissible hearsay statement) when the most accurate statement possible is the initial interview with the child. Thus, the child’s in-court testimony may only be a replication of errors developed over the course of multiple interviews by poorly trained professionals.
In an effort to determine whether a child’s statement is reliable enough to be heard in court, New Jersey imposed a pretrial “taint” procedure. The procedure was developed by the New Jersey Supreme Court in State v. Michaels, a case in which the court overturned the convictions of daycare workers who had been found guilty of sexual abuse. After reviewing interview methods used in the abuse investigation, the court determined that the convictions rested on children’s statements that were produced by suggestive interviewing. Therefore, New Jersey Supreme Court crafted new procedures to address the admissibility of child statements. Michaels requires:
[A pretrial hearing must be held] to determine whether those clearly improper interrogations so infected the ability of the children to recall the alleged abusive events that their pretrial statements and in-court testimony based on that recollection are unreliable and should not be admitted into evidence . . . The basic issue to be addressed at such a hearing is whether the pretrial events, the investigatory interviews and interrogations, were so suggestive that they give rise to a substantial likelihood of irreparably mistaken or false recollection of material facts bearing on defendant’s guilt.
The “taint hearing” is triggered when a defendant “make[s] a showing of ‘some evidence’ that the victim’s statements were the product of suggestive or coercive interview techniques.” Once the defendant establishes that sufficient evidence of unreliability exists, the burden shifts to the state to prove the reliability of the proffered statements and testimony by clear and convincing evidence. Therefore, “the ultimate determination to be made is whether . . . the totality of the circumstances surrounding the interviews, the statements or testimony retain a degree of reliability sufficient to outweigh the effects of the improper interview techniques.”
Michaels has not been uniformly followed by other states. On the contrary, where states have addressed this issue they sometimes place the burden on the party challenging the statement to demonstrate that coaching or suggestion took place. Aside from legislative determinations, a number of courts have rejected Michaels as well.
The taint procedure developed by New Jersey represents one attempt to acknowledge and address the findings of empirical research on suggestive interviewing. In light of this evidence, it is surprising that more states have not undertaken similar efforts. Even if states are unwilling to adopt a taint procedure, merely requiring judges to inquire into these issues has the potential to encourage the adoption of proper forensic interview protocols. That is, if judges require the proponents of child statements to demonstrate that the statements are not corrupted, child protection agencies might be encouraged to develop procedures that would survive such scrutiny. In the child protection context, taint hearings could be required regardless of whether the child will testify in court. As noted, because hearsay is admissible it is rarely necessary for children to appear in person.
In recent decades, advocates of children’s rights have suggested that children’s voices should be heard in legal proceedings. That view is endorsed by Article 12 of the United Nations Convention on the Rights of the Child, which requires signatories to accord children “the opportunity to be heard in any judicial and administrative proceedings.” The United States has not ratified the Convention, but an active bar of lawyers for children has done much to advance analogous goals in American courts. Advocates for increasing children’s participation in legal proceedings tend to focus on the benefits to children. Comparatively less attention has been paid to the potentially negative impact that child statements may have on the fairness of child protection proceedings. As discussed elsewhere in this article, allowing attorneys for children to participate in fact-finding can make children more believable, even where they are not testifying accurately.
Some of the objectives sought by advocates of providing children with attorneys may be achieved without allowing these attorneys to participate in the factfinding stage of dependency trial. Without an attorney’s participation in factfinding, she can: make her child clients aware of their legal options, tell them what will happen next in their case, evaluate the likelihood of prevailing, and advocate for services on their behalf. Without an attorney’s participation in factfinding, a child will still have the “right to empowerment through participation” and may benefit from “the therapeutic nature of the attorney-client relationship.” Granted, some objectives articulated by those who advocate providing representation in child welfare cases might require an attorney’s participation in factfinding. For example, some scholars argue that lawyers for children must participate in factfinding so that they can ensure that the state “meet[s] its legal burden when attempting to . . . remov[e a] child from his home or terminat[e his] parental rights.” They argue that “[o]ther parties . . . have interests and motivations other than [that of the child] and cannot adequately represent the child.” While that last point is almost certainly true, the judge or fact-finder, and not the parties, will also be present and interested in finding the truth. Because attorney participation in factfinding, prepping child witnesses and eliciting testimony creates risk of injecting unreliability into child welfare proceedings, limiting the role of attorneys for children may be a proper solution.
The empirical research discussed in this article suggests that the burden of proof in cases of child abuse and neglect should be higher than a preponderance of the evidence.
In Santosky v. Kramer, the United States Supreme Court found that the termination of parental rights requires the state establish that termination is proper by clear and convincing evidence. The Court found that the “fair preponderance” standard employed by the state in parental termination proceedings violated the Fourteenth Amendment’s Due Process Clause. Relying on the Mathews v. Eldridge rule, the Court evaluated the due process concern by weighing three considerations: (1) the private interests affected by the proceeding; (2) the risk of error created by the state’s chosen procedure; and (3) the countervailing governmental interest supporting use of the challenged procedure. Although abuse or neglect adjudications do not carry the permanent consequences of parental termination proceedings, these proceedings may deprive parents of their ability to be with their children for an extended period of time. As a New Jersey Superior Court explained the dilemma:
The Legislature has made the policy judgment that satisfying the slightest of our standards of proof—preponderance of the evidence—is all that is required to support a finding of abuse or neglect and to activate the court’s intervention to protect children. As a result, the potential of finding abuse or neglect by a parent when it actually did not occur is increased; but raising the standard of proof would increase the risk of not finding abuse or neglect when it actually did occur, posing an unacceptable threat to the safety of children.
Surely the state’s interest in promoting the safety of children is beyond question. However, the Supreme Court’s analysis of the proper burden of proof under the Fourteenth Amendment’s Due Process jurisprudence suggests the clear and convincing evidence should be required, even at this pre-termination stage. It is beyond the scope of this article to address the private and government interests at stake in these proceedings. However, the evidence presented in this article directly addresses the second factor relevant to the Court’s due process analysis: the risk of error created by the contested procedure.
As demonstrated by the empirical evidence discussed above, the risk of error in these proceedings is significant. Even if interview techniques are improved, the evidence shows that adults have a strong tendency to judge children’s statements as correct. This truth bias impacts the integrity of every stage in the investigation of a child welfare case. In light of this research, a low burden of proof in child welfare cases is particularly concerning.
There is also a symbolic value in establishing a clear and convincing evidence standard. As the Court in Santosky acknowledged with respect to termination proceedings, “[i]ncreasing the burden of proof is one way to impress the factfinder with the importance of the decision and thereby perhaps to reduce the chances that inappropriate terminations will be ordered.” An analogous symbolic effect could be helpful in pre-termination stages of child welfare proceedings—proceedings which implicate the same fundamental right to parent. Signaling to judges that they should weigh their decisions carefully could set an important tone in child welfare cases.
As this article has demonstrated, empirical research should give investigators and judges pause when they rely on a child’s account of her past experiences. This is not because children are willfully bad or manipulative—and it is not because children have bad memories. Rather, there are important developmental factors that impact a child’s ability to report her past experiences, social factors that impact an adult’s ability to discern the veracity of child testimony, and structural factors that exacerbate both of these problems. This article does not suggest that the statements of children should be disregarded. Excluding children’s accounts from the child welfare process creates an intolerable risk that abuse may go undetected. Research shows that children are sometimes reluctant to describe abuse they have suffered and may minimize or deny abuse that actually happened. In particular, children may be less likely to disclose abuse the more closely related they are to the perpetrator. This is especially problematic for family court cases. Additionally, instructions to keep a secret can have a powerful effect on some children. Given these constraints, delayed or inconsistent reports are not necessarily inaccurate. Furthermore, because many forms of child sexual abuse can occur in private and leave no physical trace, the statement of a child may, in fact, be some of the only probative evidence.
Although it has been argued that “[t]he costs of allowing an abused child to fall through the cracks are equal to the costs of wrongfully convicting an innocent person,” to adequately weigh these harms policymakers must consider the true impact of false-positives. First, poor and minority communities are disproportionately affected. This can create a stratified system in which already rampant inequality translates into decreased respect for the privacy and dignity of some targeted communities. Second, foster care is an unsafe place for children. As others have reported, “children’s subsequent treatment in the child welfare system often constitutes abuse and neglect of its own.” Third, even if foster care were safe and nurturing, children and parents have a fundamental right to be free from separation. Because of delays in the system,
it can take years to get a child out of foster care. In at least seventy-five percent of states, the majority of children who entered foster care for the first time in 2009 were still in foster care twelve months later.
There are no easy solutions to achieving a balance between including statements that may help protect children, on the one hand, and excluding statements of questionable accuracy, on the other. What is striking is the little effort the current legal regime makes to strike any kind of balance at all.
Society should not tolerate the abuse of children. But neither should society tolerate needlessly tearing families apart. Further research would help identify ways to better address both concerns.
This article has attempted to demonstrate the costs of the current child welfare system’s treatment of children’s statements. The current regime in most states suggests that courts are relying on children’s statements when they should not be. I have argued that caring about children and their safety does not necessarily mean that determinations in the child welfare context should turn on what children say. Rather, the better course for children may be to create procedures that facilitate accurate statements. In any event, important child welfare determinations should be made using the best available information. Courts and legislatures must respond to empirical research demonstrating that the child welfare system is not doing enough to ensure that this objective is achieved.
† Visiting Assistant Professor of Lawyering Skills, Seattle University School of Law. The author gratefully acknowledges the support of the Brooklyn Law School Summer Research Stipend program. Special thanks are owed to: Genevieve Urs, Henry Hwang, Amy Mulzer, Lauren Shapiro, and Jessica Marcus. The author welcomes comments at firstname.lastname@example.org.
. Kennedy v. Louisiana, 554 U.S. 407, 443 (2008).
. The United States Supreme Court has long recognized that a parent has a constitutionally protected fundamental right to raise his or her children. E.g., Troxel v. Granville, 530 U.S. 57, 66 (2000) (noting that, in light of precedent, “it cannot now be doubted that the Due Process Clause of the Fourteenth Amendment protects the fundamental right of parents to make decisions concerning the care, custody, and control of their children.”); Wisconsin v. Yoder, 406 U.S. 205, 232 (1972) (“The history and culture of Western civilization reflect a strong tradition of parental concern for the nurture and upbringing of their children. This primary role of the parents in the upbringing of their children is now established beyond debate as an enduring American tradition.”).
. As a preliminary matter, it is important to distinguish between the types of statements a child may make. Children may make statements about their past experiences, or they may express their desires for the future. For example, a child may make a statement about a past experience of abuse. On the other hand, the same child may make a statement about the family she wishes to live with or where she wishes to go to school. Those two types of statements raise very different concerns. This article is concerned primarily with the first type of statement: a child’s account of her past experiences. Another point to keep in mind is that advocates for children might argue that a child has a right to make such statements on her own behalf.
. Concern that the legal system may rely on inaccurate statements by children is hardly a new phenomenon. Notably, in the early nineteen hundreds, Alfred Binet, French psychologist and father of the modern IQ test, “recommended to French judges that they take responsibility for obtaining reliable testimony from children, and he cautioned them about the dangers of forcing their own questions into a child’s memory.” Steven Ceci & Maggie Bruck, Jeopardy in the Courtroom: A Scientific Analysis of Children’s Testimony 55 (1995) [hereinafter Ceci & Bruck, Jeopardy in the Courtroom]. Nevertheless, as discussed further below, the amount of new empirical data amassed by Bruck and Ceci, Michael Lamb, Victoria Talwar and others, provides important new lessons for the legal system.
. Due to the nature of these cases, and the way they are investigated, it is likely that child welfare determinations commonly rely on children’s statements. There is no data available on the number of calls to state child abuse registries that result from a child as the source. However, case law reviewed in this paper suggests that children’s statements do trigger child welfare investigations. In addition, children are often interviewed as part of the investigation. See Theodore Cross & Cecelia Casaneuva, Caseworker Judgments and Substantiation, 14 Child Maltreatment 38, 39 (2009) (describing interviews with the child as standardized component of caseworker investigations).
. Cross & Casaneuva, supra note 10, at 39. See, e.g., Ala. Admin. Code r. 660-5-34.04 (2007) (requiring “in person initial contact . . . be made with the children who are allegedly abused or neglected . . . and all other children in the home”); D.C. Code § 44-1301 (2001) (requiring investigation within 24 hours which includes “seeing the child and all other children in the household outside the presence of the caretaker or caregivers”); 325 Ill. Comp. Stat. 5/7.4(a) (1980) (requiring caseworkers to provide notice to families after seeing to the “safety of the child and other family members and the risk of the subsequent abuse or neglect”); Iowa Code § 232.71B(4)(a) (2011) (requiring assessment including “[i]dentification of the nature, extent, and cause of the injuries, if any, to the child named in the report”); 922 Ky. Admin. Regs. 1:330 (2010) (noting that if the report indicates non-imminent danger of physical abuse “efforts shall be made to have face-to-face contact with the child within 24 hours” and in other cases face-to-face contact should be made within 48 hours); Md. Code Ann., Fam. Law § 5-706 (West 2012) (requiring investigators “see the child” within five days of allegation of neglect and twenty-four hours after receiving report of physical or sexual abuse); Minn. Stat. § 626.556(10)(i) (2013) (noting that “[u]pon receipt of a report, the local child welfare agency shall conduct a face-to-face contact with the child reported to be maltreated . . .”); N.Y. Soc. Serv. Law § 424(6)(a) (McKinney 1992) (noting that “after seeing to the safety of the children” child protective service shall provide some notification to the subjects of the report and other persons named in the report); 23 Pa. Cons. Stat. § 6368(a) (2007) (noting that “[u]pon receipt of each report of suspected child abuse, the county agency shall immediately commence an appropriate investigation and see the child immediately if emergency protective custody is required . . . or if it cannot be determined from the report whether emergency protective custody is needed”); W. Va. Code 49-6A-9(b)(3) (2008) (requiring that each local child protective service shall, within fourteen days, conduct “face-to-face interview with the child or children”).
. Lassiter v. Dep’t of Soc. Servs. of Durham Cnty., N.C., 452 U.S. 18, 30–31 (1981) (holding that indigent parent has no due process right to appointed counsel in termination proceedings but noting that many states require appointed counsel not only in parental termination proceedings, but also in court proceedings in dependency and neglect cases as well).
. Not every state has a notice requirement, and states that do require some notice be provided to parents under investigation differ greatly in the information that must be provided. See, e.g., 325 Ill. Comp. Stat. 5/7.4(b)(3) (1980) (“the Department shall forthwith notify the subjects of the report in writing, of the existence of the report and their rights existing under this Act in regard to amendment or expungement”); N.H. Rev. Stat. Ann. § 169-C:34(VI) (2013) (“At the first contact in person, any person investigating a report of abuse or neglect on behalf of the department shall verbally inform the parents of a child suspected of being a victim of abuse or neglect of the specific nature of the charges and that they are under no obligation to allow a social worker or state employee on their premises or surrender their children to interviews unless that social worker or state employee is in possession of a court order to that effect.”); S.C. Code Ann. § 63-7-920 (2008) (“The department must furnish to parents or guardians on a standardized form [several pieces of] information as soon as reasonably possible after commencing the investigation [including t]he allegations being investigated . . . whether the person’s name has been recorded by the department as a suspected perpetrator of abuse or neglect . . . [and] how information provided by the parent may be used, the possible outcomes of the investigation”); Tenn. Code Ann. § 37-1-406(a) (2013) (“All representatives of the child protective services agency shall, at the initial time of contact with the individual who is subject to a child abuse and neglect investigation, advise the individual of the complaints or allegations made against the individual consistent with laws protecting the rights of the informant.”); Va. Code Ann. § 63.2-1516.01 (2012) (“The local department shall, at the initial time of contact with the person subject to a child abuse and neglect investigation, advise such person of the complaints or allegations made against the person, in a manner that is consistent with laws protecting the rights of the person making the report or complaint.”).
. There is some variation in the terminology employed by states. For example New York and Illinois use the term “indicated” to refer to an unfavorable disposition, where other states use the terms founded/unfounded or substantiated/unsubstantiated.
. These numbers raise a variety of questions beyond the scope of this article. Whether these services are warranted, and whether families view these referrals as voluntary or coerced are outside the bounds of this paper.
. See Martin Guggenheim, Somebody’s Children: Sustaining the Family’s Place in Child Welfare Policy: Nobody’s Children: Abuse and Neglect, Foster Drift, and the Adoption Alternative. by Elizabeth Bartholet. Boston: Beacon Press. 1999. Pp. VIII, 3, 113 Harv. L. Rev. 1716, 1725 (2000) (discussing study which found that ninety percent of child welfare caseload in one jurisdiction “involve[d] less serious physical abuse (for example, a single, minor injury such as a bruise or a scratch) or less severe neglect (such as parental drug or alcohol abuse with no other apparent protective issues, dirty clothes or a dirty home, lack of supervision of a school-age child, or missed school or medical appointments)…”). As Professor Guggenheim describes, these lower-risk neglect cases are often poverty-related. The trend towards investigating and prosecuting low-risk parenting infractions has not abated in recent years. E.g., In re Arthur S., 68 A.D.3d 1123, 1124 (N.Y. App. Div. 2009) (noting that the presumption of child neglect created by marijuana use is not rebutted by a showing that “the children were never in danger and were always well kept, clean, well fed and not at risk”).
. Child Maltreatment 2011, supra note 2, at 17 (“[M]ore than 3.7 million (duplicate count) children were the subjects of at least one report. One-fifth of these children were found to be victims with dispositions of substantiated (18.5%), indicated (1.0%), and alternative response victim (0.5%). The remaining four-fifths of the children were found to be non-victims of maltreatment.”).
. See Child Maltreatment 2011, supra note 2, at 29. (This cohort was divided (18.5% substantiated, 1.0% indicated, 0.5% alternative response victim, 9.3% alternative response nonvictim, 58.9% unsubstantiated, 0.1% intentionally false, 1.8% closed with no finding, 9.1% no alleged maltreatment, 0.8% other and 0.1% unknown)).
. Cross & Casaneuva, supra note 10, at 41 (describing conclusions of another study which found that “the verbal formula”—the legal standard—hardly affects the caseworkers decision to substantiate a case).
. Shani King, The Family Law Canon in a (Post?) Racial Era, 72 Ohio St. L.J. 575, 603 (2011) (quoting Dorothy Roberts, Shattered Bonds: The Color of Child Welfare 8 (2002)). The reasons for this disparity have been debated elsewhere. See Andrea J. Sedlak, Jane Mettenburg, Monica Basena, Ian Petta, Karla McPherson, Angela Green, & Spencer Li, Fourth National Incidence Study of Child Abuse and Neglect 9 (2010); Elizabeth Bartholet, The Racial Disproportionality Movement in Child Welfare: False Facts and Dangerous Directions, 51 Ariz. L. Rev. 871, 871 (2009) (claiming that “[t]he evidence indicates that black children are in fact disproportionately victimized by maltreatment”). Although this paper does not directly address the question of racial disparity in child maltreatment, it does shed light on problems with using substantiation rates as an actual measure of child maltreatment. See Cross & Casaneuva, supra note 10, at 41.
. N.Y. State Office of Children and Family Servs., Feasibility Study of Family and Supreme Court Access to the Statewide Central Register of Child Abuse and Maltreatment, Interim Report to the Governor and New York State Legislature, 9 (2009), available at http://www.ocfs.state.ny.us/main/reports/Feasibility%20Study%20Interim
%20Report.pdf). According to Census Bureau population estimates, New York City’s population increased from 8,175,133 in April of 2010 to 8,405,837 in July of 2013. City of New York, Department of City Planning, Current Estimates of New York City’s Population for July 2013, available at http://www.nyc.gov/html/dcp/html/census/popcur.shtml.
. U.S. Dep’t of Health & Human Servs., Interim Report to the Congress on the Feasibility of a National Child Abuse Registry, 29 (2009). See also Humphries v. Los Angeles County, 554 F.3d 1170, 1201 (9th Cir. 2009), rev’d in part on other grounds, 131 S. Ct. 447 (2010) (finding California’s lack of procedure, which enables parents to remove themselves from the Child Abuse Central Index is unconstitutional under the Due Process Clause).
. Dupuy v. McDonald, 141 F. Supp. 2d 1090, 1102 (N.D. Ill. 2001) (“Ultimately, 74.6% of indicated reports were expunged in the appeals process.”); Valmonte v. Bane, 18 F.3d 992, 1003 (2d Cir. 1994) (acknowledging party’s presentation of seventy-five percent reversal rate).
. Child Maltreatment 2011, supra note 2, at 83 (referring to study of forty-six states). According to the study, “[c]ourt action may include any legal action taken by the CPS agency or the courts on behalf of the child, including authorization to place a child in foster care and filing for temporary custody, protective custody, dependency, or termination of parental rights.” Id.
. Id. (finding that “[a]nalyzing data from the States that report both foster care and in-home postresponse services reveals that three-fifths (62.6%) of victims (duplicate count) who received postresponse services received only in-home services”).
. The burden of proof for termination of parental rights is higher, a clear and convincing evidence standard. Santosky v. Kramer, 455 U.S. 745, 747–48 (1982). The Indian Child Welfare Act of 1978 permits termination of parental rights only upon “evidence beyond a reasonable doubt” Indian Child Welfare Act, 25 U.S.C. § 1912(f) (2012).
. E.g., N.Y. Fam. Ct. Act §1031(f) (McKinney 2010); In re Archer, 744 N.W.2d 1, 9 (Mich. Ct. App. 2007) (“Circumstances indicating the reliability of a hearsay statement may include spontaneity, consistent repetition, the mental state of the declarant, use of terminology unexpected of a child of similar age, and lack of motive to fabricate.”); State v. Boston, 545 N.E.2d 1220 (Ohio 1989) (expanding type of evidence allowable and unobjectionable in child abuse cases).
. E.g., In re Blaize F., 855 N.Y.S.2d 284, 286 (App. Div. 2008) (crediting statements of child that contradicted the statements of step-father and contradicted account of event she had previously given her mother); In re Kayla J., 903 N.Y.S.2d 601, 603 (App. Div. 2010) (noting that “[r]esolution of this very difficult case turned almost entirely on issues of credibility.”); In re R.M., 718 N.E.2d 550, 556 (Ill. App. Ct. 1999) (crediting child’s accusations of cigarette burns, possibly on regular basis, by parents despite reluctance and delay in reporting this and despite lack of reliable physical indications of past burns); In re Welfare of Child of C.A.P. & E.E.G., No. A09–922, 2009 WL 3364308, at *2 (Minn. Ct. App. 2009) (enterin finding of abuse after finding child credible though she appeared “distracted and reluctant to answer several questions” while testifying, the child’s mother believed she was lying, and family friends and child’s brother testified that they did not observe indications of the inappropriate contact alleged by child).
. See, e.g., Guadalupe A. v. Superior Court, 285 Cal. Rptr. 570 (Ct. App. 1991) (“When ruling in dependency proceedings, the welfare of the minor is the paramount concern of the court. The purpose of these proceedings is not to punish the parent, but to protect the child. . . . Consequently, a trial court should not restrict or prevent testimony on formalistic grounds. On the contrary, the court should avail itself of all evidence which might bear on the child’s best interest.”).
. See Elisa Krackow & Steven J. Lynn, Is There Touch in the Game of Twister? The Effects of Innocuous Touch and Suggestive Questions on Children’s Eyewitness Memory, 27 L. & Hum. Behav. 589 (2003).
. See In re Charlie S., 920 N.Y.S.2d 187, 188 (App. Div. 2011) (“[T]he child’s out-of-court statements alleging that the father inappropriately touched his buttocks were sufficiently corroborated by testimony from the child’s case worker and from his high school principal, both of whom stated that the child related to them that such activity occurred.”).
. The dispositional hearing of a child welfare case is akin to the sentencing portion of a criminal trial. Following a finding of abuse or neglect, the court issues a disposition: an order against the parent requiring certain remedial measures. E.g., In re Madison T., 153 Cal. Rptr. 3d 437, 440 (Ct. App. 2013) (“In general, a social worker’s testimony at a dispositional hearing is largely based on hearsay. If it is appropriate for the social worker to rely on a particular item of evidence in making a dispositional recommendation, the fact that the evidence is hearsay is not
. See Coleman, supra note 19, at 539 (“The states’ view of the children’s own interests is
consistent with their public policy objectives. That is, they assume the children would want state officials to take no chances, and to leave no stone unturned, in their effort to substantiate a maltreatment report.”).
. See Kelly L.Warren, Elyse Dodd, Graham Raynor, & Carole Peterson, Detecting Children’s Lies: Comparing True Accounts About Highly Stressful Injuries with Unprepared, Prepared and Coached Lies, 30 Behav. Sci. & L., 329, 329 (2012) (citing studies showing that children have elaborate memories about events such as personal injuries, natural disasters, and painful medical procedures). But see David J. La Rooy, Lindsay C. Malloy, & Michael E. Lamb, The Development of Memory in Childhood, in Children’s Testimony: A Handbook of Psychological Research and Forensic Practice 49, 60 (Michael E. Lamb, David J. La Rooy, Lindsay C. Malloy, & Carmit Katz, eds., 2nd ed. 2011) [hereinafter La Rooy, Malloy & Lamb, The Development of Memory in Childhood].
The relations between stress and memory are inconsistent, with some studies indicating enhanced performance by children and others showing detrimental effects of stress. Importantly, however, memories of stressful or even traumatic experiences are subject to the same encoding, storage, and retrieval principles as are memories for more mundane events, meaning that we can forget traumatic events, just as we can forget other experiences, that traumatic or stressful experiences are not necessarily remembered in richer detail just because the events were traumatic and that all memories can be contaminated.
. David J. La Rooy, Lindsay C. Malloy, & Michael E. Lamb, Setting Realistic Expectations: Developmental Characteristics, Capacities and Limitations, in Children’s Testimony: A Handbook of Psychological Research and Forensic Practice 22, 23 (Michael E. Lamb, David J. La Rooy, Lindsay C. Malloy, & Carmit Katz, eds., 2nd ed. 2011) [hereinafter La Rooy, Malloy & Lamb, Setting Realistic Expectations].
. See Ceci & Bruck, Jeopardy in the Courtroom, supra note 8, at 77–78. Stephen Ceci recounts a conversation between a mother and a four year old child that captures the concern:
“Mother: So tell me about his crayon.
Child: It’s a special crayon.
Child: And sparks.
Mother: What do you mean sparks?
Child: Sparks come out of the crayon.
Mother: When you draw, you mean?
Mother: Oh, wow. You mean like fire sparks?
Child: Ya sparks.”
Ceci writes, “[t]he child, who was a subject in one of our experiments, was trying to tell his mother about a crayon that has ‘sparkles.’ Having never seen that type of crayon before, the mother inaccurately concluded that the crayon burns a hole in the paper.”
. See Clyde Haberman, The Trial That Unleashed Hysteria Over Child Abuse, N.Y. Times, Mar. 10, 2014; State v. Michaels, 642 A.2d 1372 (N.J. 1994); Michael E. Lamb, Yael Orbach, Irit Hershkowitz, Phillip W. Esplin, & Dvora Horowitz, Structured Forensic Interview Protocols Improve the Quality and Informativeness of Investigative Interviews with Children: A Review of Research Using the NICHD Investigative Interview Protocol, 31 Child Abuse & Neglect 1201, 1202 (2007) [hereinafter Lamb, Orbach, Hershkowitz, Esplin & Horowitz, Structured Forensic Interview Protocols] (noting that widespread publicity about “infamous cases” led many researchers to study children’s capacity to provide accurate accounts of their past experiences).
. E.g., State v. Michael H., 970 A.2d 113, 122 (Conn. 2009) (noting “the defendant has failed to make any showing that such testimony was the product of unduly coercive or suggestive questioning”); State v. Bumgarner, 184 P.3d 1143, 1150 (Or. Ct. App. 2008) (discussing “taint hearings” as a way to deal with “coercive child interviewing”); Commonwealth v. Davis, 939 A.2d 905, 910 (Pa. Super. Ct. 2007) (noting that “the suggestive technique and content of the interviews provided clear and convincing evidence that J.D.’s later recollections were tainted and a product of coercion, not of his own memory”).
. See Simona Ghetti & Kristen Weede Alexander, “If It Happened, I Would Remember It”: Strategic Use of Event Memorability in the Rejection of False Autobiographical Events, 75 Child Dev. 542, 542 (2004) (“Numerous studies have replicated the finding that postevent misleading information may distort children’s memory of details of an event, and that sizable percentages of children and adults may at times create memories for entirely false events.”); J. Zoe Klemfuss & Stephen Ceci, Normative Memory Development and the Child Witness, in The Evaluation of Child Sexual Abuse Allegations 153, 172 (Kathryn Kuehnle & Mary Connell eds., 2009) (“There is evidence that children who have been interviewed suggestively often come to believe the suggestions because the false information takes the place of the actual event details in memory.”).
. See The Suggestibility of Children’s Memory, supra note 76, at 423 (“Interviewer bias characterizes those interviewers who hold prior beliefs about the occurrence of certain events and, as a result, mold the interview to maximize disclosures from the interviewee that are consistent with the interviewer’s prior beliefs.”).
. In another study, Bruck and Ceci found: “when the interviewer was biased in a direction that contradicted the activity viewed by the child, those children’s stories conformed to the suggestions or beliefs of the interviewer.” Bruck & Ceci, The Suggestibility of Children’s Memory, supra note 76, at 424–25. Then, when later asked neutral questions by their parents, the children’s answers remained consistent with the interviewers’ biases. Id. at 425. See also Tamar R. Birckhead, The Age of the Child: Interrogating Juveniles After Roper v. Simmons, 65 Wash. & Lee L. Rev. 385, 417 (2008) (noting that “juveniles are particularly susceptible to pressure from authority figures” and that “they are more compliant and open to suggestion, repetition, and other social influence tactics than adults”).
. Id. Maggie Bruck & Steven Ceci, Children’s Suggestibility, in Memory for Everyday and Emotional Events 382 (Nancy L. Stein, Peter A. Ornstein, Barbara Tversky, Charles Brainerd eds., 1997) (“When interviewing children, adults frequently repeat a question because the child’s first response may not provide enough information. In forensic interviews, questions may be repeated to check the consistency of a child’s reports.”).
. Id. at 119 (“A number of studies, from different domains, demonstrate that when young children are asked the same question more than once within an interview, they change their answer.”). See also Bruck & Ceci, The Suggestibility of Children’s Memory, supra note 76, at
. For example, one study compared preschoolers’ responses to questions that contained a definite article (“the”) or quantifier (“some v. any”) with questions that contained an indefinite article. When participants were asked, for instance, “did you see the hat?” they responded affirmatively more frequently then when they were asked questions that included the indefinite article “did you see a hat?” Krackow & Lynn, supra note 62, at 591, 597 (“Our results indicated that compared with children who are asked direct questions, children who are asked suggestive ‘tag’ questions on a single occasion were more likely to falsely assent to questions that may provoke concerns in a child abuse investigation and in the legal arena.”).
. In one study, children visited their pediatricians when they were five years old; during that visit a male pediatrician gave each child a physical exam, an oral polio vaccine and an inoculation. During the same visit a female research assistant talked to the child about a poster on the wall, read the child a story, and gave the child some treats. The children were interviewed a year later, some children were reminded of the visit but given wrong information about what happened. Other children were given no misinformation. Children who were given no misinformation gave highly accurate final reports. Children who were given misleading reminders gave highly inaccurate reports, with half of the children giving inaccurate information along the lines of what was suggested and thirty-eight percent of those children including other non-suggested and inaccurate information. Meaning, even if they were only reminded that the female pediatrician gave them the inoculation and oral vaccine they also reported that female had checked their ears and nose. Ceci & Bruck, Jeopardy in the Courtroom, supra note 8, at 111.
. In any legal proceeding involving a child witness, numerous interviews with a child are practically inevitable: even before child welfare or other law enforcement personnel become involved in a case, that child may have spoken to parents, siblings, neighbors, school personnel or therapists. Ceci & Bruck, Jeopardy in the Courtroom, supra note 8, at 80, 107. Once a case is brought to the attention of authorities, experts speculate that child victims of family violence are subjected to at least a dozen investigative interviews before legal proceedings can be resolved. Id. at 80.
. See Angela D. Evans and Kang Lee, Promising to Tell the Truth Makes 8- to 16-year-olds More Honest, 28(6) Behavioral Sci Law. 801, 810 (2010) (noting that, consistent with earlier research involving younger children, majority of 8- to 16-year-olds in their study cheated on trivia test when provided easy access to answers and then lied about it).
. Id. at 801 (describing “developmental increase in lie-telling” and noting existing research which found that children begin to tell lies during preschool years, that by four years of age children demonstrate clear signs of intentions to deceive others, that children’s lie-telling behavior also appears to increase with age, that four-year-olds tend to lie more than two-year-olds, and that seven- and eight-year-olds tended to tell lies more frequently than six-year-olds).
. Victoria Talwar, Kang Lee, Nicholas Bala, & R. C. L. Lindsay, Children’s Lie-Telling to Conceal a Parent’s Transgression: Legal Implications, 28 Law & Hum. Behav., 411, 412 (2004) [hereinafter Talwar, Lee, Bala & Lindsay, Legal Implications].
. Victoria Talwar, Heidi M. Gordon, & Kang Lee, Lying in the Elementary School Years: Verbal Deception and Its Relation to Second-Order Belief Understanding, 43(3) Dev. Psychol. 804, 810 (2007).
. Eitan Elaad, Shiri Lavy, Diana Cohenca, Ester Berholz, Pnina Thee, & Yaara Ben-Gigi., Lies, Truths, and Attachment Orientations in Late Adolescence, 52 Personality & Individual Differences 670, 670–73 (2012).
. For example, some studies have attempted to identify the “false-positive” rate for allegations of sex abuse and child sex abuse. Some researchers estimate that as many as twenty percent of such allegations could be false. Kristine A. Peace, Stephen Porter & Daniel Almon, Sidetracked by Emotion: Observers’ Ability to Discriminate Genuine and Fabricated Sexual Assault Allegations, 17 Legal & Criminological Psychol. 322, 323 (2011). Yet, these studies may rely on a caseworker’s decision to substantiate the report as support for the conclusion that the allegation was true. One implication of the research here is that caseworker substantiation may not be a useful measure of the truth of a child’s statement. See Everson & Boat, supra note 39, at 234 (suggesting a 4.7 to 7.6% false allegation rate for child and adolescent accusations of sexual abuse, based on caseworker substantiation rates).
. Id. at 872–73. This study of 123 children showed that, on average, children peeked eleven seconds after the experimenter left the room. Half the children peeked within six seconds. Age and gender were not significantly correlated with behavior.
. Talwar, Lee, Bala & Lindsay, Legal Implications, supra note 116, at 428. See also Kari L. Nysse-Carris, Bette L. Bottoms, & Jessica M. Salerno, Experts’ and Novices’ Abilities to Detect Children’s High-Stakes Lies of Omission, 17 Psychol. Pub. Pol’y & L. 76, 79 (2002) (finding that children, especially older (five- to six-year-old) versus younger (three- to four-year-old) children, will sometimes lie to protect their parent who has broken a toy and asked them to lie about it).
. Thomas D. Lyon, Lindsay C. Malloy, Jodi A. Quas, & Victoria A. Talwar, Coaching, Truth Induction, and Young Maltreated Children’s False Allegations and False Denials, 79 Child Dev., 914, 925 (2008) (“[C]oaching has consistent and robust effects on children’s honesty.”) [hereinafter Lyon, Malloy, Quas & Talwar, Coaching Truth Induction]. Interestingly, one study of preschool and elementary school children found that elementary school children were more susceptible to coaching than preschoolers. Id. at 922; Talwar, Lee, Bala & Lindsay, Legal Implications, supra note 116, at 413; Paul Elkman, Why Lies Fail and What Behaviors Betray a Lie, in Proceedings of the NATO Advanced Study Institute on Credibility Assessment 71, 71-81 (John C. Yuille ed. 1989).
. Lene Arnett Jensen, Jeffrey Jensen Arnett, S. Shirley Feldman, & Elizabeth Cauffman, The Right to Do Wrong: Lying to Parents Among Adolescents and Emerging Adults, 33 J. of Youth & Adolescence 101, 109 (2004).
. Researchers note that the literature “contains many references to professionals who ill-advisedly adhere to the maxim, ‘Children never lie about sex abuse.’” Everson & Boat, supra note 39, at 235.
. “Professionals cannot reliably discriminate between children whose reports are accurate from those whose reports are inaccurate as the result of suggestive interviewing techniques.” Id. But see Nysse-Carris, Bottoms, & Salerno, supra note 129, at 91 (noting that although most studies find that adults can detect child lies at chance levels, in study of high-stakes lies of omission, rate of accurate detection was significantly greater than chance) Still, even the most accurate judges were unable to detect one third of lies told by children.
. Rachel L. Laimon & Debra A. Poole, Adults Usually Believe Young Children: The Influence of Eliciting Questions and Suggestibility Presentations on Perceptions of Children’s Disclosures, 32 L. & Hum. Behav. 489 (2008).
. Lay judges believed seventy percent of fabricated reports that were coached by parents. Id. “The finding that parental coaching of children to lie is so effective has dismaying implications for the courts.” Id.
. Nysse-Carris, Bottoms, & Salerno, supra note 129, at 76. (reviewing other studies that found that federal law enforcement officers, sheriffs, and forensic psychologists detected lies about adults’ strongly held opinions with somewhat greater accuracy (sixty-seven percent accuracy or higher) than did federal judges and clinical psychologists).
. Id. See also Peace, Porter, & Almon, supra note 119, at 330 (finding inverse relationship between confidence and detection accuracy: “participants who felt more confident in their detection abilities (overall) also believed they would be more successful in their veracity determinations . . . whereas their actual performance was negatively related to their confidence”).
. Jonathan M. Golding, Emily E. Dunlap, & Emily C. Hodell, Jurors’ Perceptions of Children’s Eyewitness Testimony, in Children as Victims, Witnesses, and Offenders: Psychological Science and the Law 200 (Bette L. Bottoms, Cynthia J. Najdowski, & Gail S. Goodman eds. 2009).
. Warren, Dodd, Raynor, & Peterson, supra note 67, at 330 (“[P]eople base their judgments in part on visual cues suggesting nervousness, agitation, or avoidance. However these cues are not reliable indicators of deception for either adults or children.”).
. Lindholm, supra note 157, at 1301. Of course, this flies in the face of the dominant view in American jurisprudence which holds that fact-finders must be able to evaluate the credibility of a witness by judging their in-person demeanor. Coy v. Iowa, 487 U.S. 1012, 1019 (1988) (“It is always more difficult to tell a lie about a person ‘to his face’ than ‘behind his back.’ In the former context, even if the lie is told, it will often be told less convincingly.”); N.J. Div. of Youth & Family Servs. v. J.F., No. FN-13-213-10, 2012 WL 3116765, at *5 (N.J. App. Div. July 20, 2012) (noting that “[w]e give deference to the trial court’s factual findings based on the trial judge’s familiarity with the case, opportunity to make credibility judgments based on live testimony, and expertise in family and child welfare matters”).
. For example, in one study caseworkers were asked to evaluate the reasons why an abuse report was not credible. Fifty-nine percent indicated that “the abuse report was…a deliberate fabrication by the adolescent or child for secondary gain.” Id.
. See, e.g., D. DePanfilis, & M. K. Salus, Office on Child Abuse and Neglect, Child Protective Services: A Guide for Caseworkers (2003), available at https://www.childwelfare.gov/pubs/usermanuals/cps/cpsf.cfm#interview (noting that in well-handled investigation the worker will involve children during “exploration of the allegations”).
. There has been a trend towards the creation of child advocacy centers where better-trained, multi-disciplinary teams collaborate to interview the alleged child victim. Jerome R. Kolbo & Edith Strong, Multidisciplinary Team Approaches to the Investigation and Resolution of Child Abuse and Neglect: A National Survey, 2 Child Maltreatment 61, 67 (1997). These interviews can become the core of a case against a parent. See infra Part V.a.
. Lamb, Orbach, Hershkowitz, Esplin & Horowitz, Structured Forensic Interview Protocols, supra note 74. Accord Nancy E. Walker, Forensic Interviews of Children: The Components of Scientific Validity and Legal Admissibility, 65 L. & Contemp. Probs. 149 (2002) (discussing, among other things, necessity of establishing and maintaining standards for quality control in conducting and evaluating forensic interviews of children); Michael E. Lamb & Deirdre A. Brown, Conversational Apprentices: Helping Children Become Competent Informants About Their Own Experiences, 24 Brit. J. of Developmental Psychol. 215 (2006) (reviewing interview techniques that either enhance or decrease children’s recall performance); Kathleen Coulborn Faller, Interview Structure, Protocol, and Guidelines, in Interviewing Children About Sexual Abuse: Controversies and Best Practice 66, 66–67 (Kathleen Coulborn Faller ed., 2007).
. Michael E. Lamb, Kathleen J. Sternberg, Yael Orbach, Phillip W. Esplin, & Susanne Mitchell, Is Ongoing Feedback Necessary to Maintain the Quality of Investigative Interviews with Allegedly Abused Children?, 6 Applied Developmental Sci. 35, 35 (2002).
. Some of the national protocols include: American Professional Society on the Abuse of Children (APSAC) Guidelines for Practice APSAC (a “narrative interview” approach with an emphasis on research-based free recall techniques aimed at eliciting reliable verbal narratives from children), www.apsac.org; Cornerhouse Forensic Interview Model (a semi-structured interview process, following a RATAC protocol: Rapport, Anatomy identification, Touch survey, Abuse scenario, and Closure), www.cornerhousemn.org/index.html; National Children’s Advocacy Center (NCAC) Child Forensic Interview Structure (a flexible interview structure with a two-stage approach), www.nationalcac.org; National Institute of Child Health and Human Development (NICHD) Protocol: The Ten Step Investigation Interview (an interviewer training approach developed with reference to childhood development issues), http://nichdprotocol.com/.
. Mich. Comp. Laws § 722.628(6) (2008). But see In re Showers, 255007, 2004 WL 2314965 (Mich. Ct. App. Oct. 14, 2004) (noting “respondent has offered no authority in support of the proposition that as a prerequisite to the admissibility of a child’s statements concerning sexual abuse under MCR 3.972(C)(2), the interviewers who obtain the statements must have followed a forensic interview protocol”). Some states require specialized forensic training when the investigation involves a serious offense. E.g., La. Child. Code Ann. art. 508(B) (2003); Minn. Stat. Ann. § 626.556 (10)(i) (2013).
. A number of states have adopted the RATAC protocol taught by CornerHouse, a private agency referenced above. The staff at CornerHouse only conduct interviews when a child is alleged to be a victim of sexual assault or to have witnessed a violent crime. But, the program also teaches others, including state caseworkers, to implement their interview protocol. While the interview protocol is only semi-structured, it requires intense training to become proficient as a RATAC interviewer. Jennifer Anderson, Julie Ellefson, Jodi Lashley, Anne Lukas Miller, Sara Olinger, Amy Russell, Julie Stauffer, & Judy Weigman, The Cornerhouse Forensic Interview Protocol: RATAC, 12 T.M. Cooley J. Prac. & Clinical L. 193 (2009).
. See, e.g., Child Maltreatment 2011, supra note 2, at 138, 147, 172, 195, 221. Although technically New York has a forensic interview protocol, based on conversations with New York family court practitioners, in practice caseworkers do not rely on the protocol.
. See id. at Table 3-1. Interestingly, the fifth state, Michigan, is the only state that requires a forensic interview by law. Mich. Comp. Laws § 722.628 (2008) (Michigan substantiated cases at about the same rate (21%) as California (23%), Florida (18%), and Texas (28%). New York substantiated 43% of cases. California, Florida, Michigan, and Texas all require a preponderance of the evidence to substantiate a report of child maltreatment, where New York only requires some credible evidence).
. Researchers note that, “[g]enerally speaking, there was little effort directed at instructing children on the ground rules of interviewing, building rapport with the children, and practicing narrative responses before moving to the topic of interest, despite empirical support showing these techniques improve children reports.” Lyon & Saywitz, supra note 4, at 839.
. Livia L. Gilstrap, A Missing Link in Suggestibility Research: What is Known About the Behavior of Field Interviewers in Unstructured Interviews with Young Children?, 10 J. of Experimental Psych.: Applied 13, 15 (2004).
. Frontline, Child Welfare System FAQ, Public Broadcasting Service, http://www.pbs.
org/wgbh/pages/frontline/shows/fostercare/inside/welfarefaq.html (last visited Oct. 20, 2013) (citing caseworker educational credential data; also citing states that provide “one day’s to several days’ worth” of on-the-job training). But see, e.g., Miss. Code. Ann. § 43-1-55 (2013).
. Lamb, Sternberg, Orbach, Esplin, & Mitchell, supra note 178, at 40 (explaining that “many workshops and training programs have been designed to improve adherence to professionally endorsed practices. Unfortunately, training programs of this sort typically have little impact on the investigative techniques employed by forensic investigators.”).
. See, e.g., State v. Cain, 427 N.W.2d 5, 9 (Minn. Ct. App. 1988) (denying additional forensic evaluation where expert psychologist expressed concern about suggestive and reinforcing methods by state’s experts, and proposed to interview children only by nonintrusive, “free recall” method of evaluation, where Court found “[i]t is unclear how such an examination could ‘produce a maximum reliable statement,’ on whether abuse occurred.”); In re Adams, 308982, 2012 WL 6217033 at *5 (Mich. Ct. App. Dec. 13, 2012) (noting, “[r]espondent-mother further argues that the trial court should not have relied on petitioner’s witnesses who interviewed the children because it was not shown that the investigators who conducted the interviews followed proper forensic protocol. However, the primary investigator who initially questioned AC testified that she was trained in forensic questioning procedures and did not ask leading questions.”); State v. Mort, 321 S.W.3d 471, 484 (Mo. Ct. App. 2010) (finding that “any variations in the protocol of the state agency and the Children’s Center were not relevant and would unnecessarily divert the attention of the jury”).
. See, e.g., In re Stephen M., W10CP02014057A, 2008 WL 5572975 at *6 (Conn. Super. Ct. Dec. 31, 2008)(“The conclusion [that the interview in question was open, neutral and objective] would have been much easier to reach if the evaluation had been videotaped.”).
. Fed. R. Evid. 801. See also Judy Yun, A Comprehensive Approach to Child Hearsay Statements in Sex Abuse Cases, 83 Colum. L. Rev. 1745, 1747 (1983) (“The rule against admission of hearsay statements stems from the long-established belief that cross-examination is the best vehicle for discovering the truth and that the most reliable statements come from the witness stand.”).
. See Crawford v. Washington, 541 U.S. 36, 54 (2004) (discussing relationship between the rule against hearsay and the Sixth Amendment’s Confrontation Clause); Lee v. Illinois, 476 U.S. 530, 543 (1986) (noting that presumption of unreliability may be overcome by “indicia of reliability” allowing for admission of hearsay evidence, notwithstanding Confrontation Clause).
. Jonathan Fly, People v. Argomaniz-Ramirez: Examining Child Hearsay and the Confrontation Clause, 28 Am. J. Trial Advoc. 727, 731–35. See also 705 Ill. Comp. Stat. 405/2-18(4)(c) (1996) (providing that “[p]revious statements made by the minor relating to any allegation of abuse or neglect shall be admissible in evidence. However, no such statement if uncorroborated and not subject to cross-examination, shall be sufficient in itself to support a finding of abuse or neglect.”); Mich. Comp. Laws § 3.972(C)(2) (2007) (any statement made by a child under the age of 10 regarding an act of sexual abuse may be admitted into evidence through the testimony of a person who heard the child make the statement “if the court has found, in a hearing held before trial, that the circumstances surrounding the giving of the statement provide adequate indicia of trustworthiness”); N.J. Stat. Ann. § 9:6-8.46(a) (West 2005); N.Y. Fam. Ct. Act § 1046(c)(vi) (McKinney 2009).
. See generally Child Welfare Information Gateway, Representation of Children in Child Abuse and Neglect Proceedings (2012), available at https://www.childwelfare.gov
. This concern about the role of attorney for children in both supporting and interviewing children is further supported by research suggesting that clinical psychologists, as opposed to forensic interviewers, will make efforts to “normalize ambivalence” and “side-step resistance” in order to build trust. Lyon & Saywitz, supra note 4, at 844.
. Talwar & Lee, supra note 120, at 878. See also Lyon & Saywitz, supra note 4, at 853 (“Judges often take it upon themselves to provide instructions to child witnesses, but tend to fall prey to the same difficulties experienced by researchers when they first started exploring the utility of instructions: judges have to be taught to provide the child with practice questions, and not to unduly encourage ‘I don’t know’ responses.”).
. Nicholas Bala, Kang Lee, Rod Lindsay, & Victoria Talwar, A Legal & Psychological Critique of the Present Approach to the Assessment of the Competence of Child Witnesses, 38 Osgoode Hall L.J. 409, 411 (2000).
. Laurie Shanks, Evaluating Children’s Competency to Testify: Developing a Rational Method to Assess a Young Child’s Capacity to Offer Reliable Testimony in Cases Alleging Child Sex Abuse, 58 Clev. St. L. Rev. 575 (2010); In re E.G., F058381, 2010 WL 2407323, at *3 (Cal. Ct. App. June 17, 2010) (describing family court competency inquiry in which child was interviewed ex parte by judge, who inquired of her if she felt “like [she] told [him] the truth today?” and she answered affirmatively).
. Frank E. Vandervort, A Search for the Truth or Trial by Ordeal: When Prosecutors Cross-Examine Adolescents How Should Courts Respond?, 16 Widener L. Rev. 335, 337 (2010) (describing experience, in which “[l]awyers were on their best behavior, and judges would tolerate not the least bit of aggressive questioning”). See also Zajac & Hayne, I Don’t Think That’s What Really Happened: The Effect of Cross-Examination on the Accuracy of Children’s Reports, 9 J. of Experimental Psychol.: Applied 187, 187–95 (2003).
. U.S. Department of Health and Human Services, Child Protective Services: a Guide for Caseworkers (2003), available at https://www.childwelfare.gov/pubs/usermanuals/cps/cpsf.cfm#interview.
. Burns v. Arkansas Dept. of Human Servs., CA 98-541, 1999 WL 258538, at *2 (Ark. Ct. App. Apr. 28, 1999). (upholding a finding of abuse that was based on child’s statements to psychologist. Psychologist felt child’s statements were credible because of consistency with which child related act of abuse. Yet concurring judge wrote that, while the psychologist “characterized the child as frightened, fearful and apprehensive in her sessions with her . . .” a video-tape of one of those sessions “depicts a poised child, calmly, and in a matter-of-fact manner answering Dr. Brunell’s inquiries, while repeatedly interrupting to ask when she would receive her ‘treat,’ suggesting that the child was offered a reward for her statements.”).
. Karen Saywitz, A Holistic Approach to Interviewing and Treating Children in the Legal System, in Child Sexual Abuse: Disclosure, Delay, and Denial 223 (Margaret-Ellen Pipe, Michael E. Lamb, Yael Orbach, & Ann-Christin Cederborg, eds., 2007).
. In re Edward T., 343 Ill. App. 3d 778, 785 (2003) (noting that after child made allegations to caseworker that he was “whooped” by his parents, he was referred to therapist, whom he told that his parents would burn his mouth with a cigarette, and that his mother stuck pins in his nose until it started to bleed); In re Stephen M., W10CP02014057A, 2008 WL 5572975, at *7 (Conn. Super. Ct. Dec. 31, 2008) (“By the end of therapy in April 2008, through a process of making Stephen comfortable and safe, Stephen was able to verbalize his history of sexual experience, control his physical behaviors, cease his public masturbation, and reduce to non-existence his maladaptive behaviors.”).
. Id. at * 1 (noting that Jeffrey was “a premature baby, was developmentally delayed, physically much smaller than other children his age, and facially ‘different’ from other children his age.” Jeffrey’s mother took him to several doctors who diagnosed him as having a medical condition known as “Dwayne’s Syndrome,” which accounted for his facial differences).
. Id. at * 6. Pursuant to the case plan, mother, stepfather and father all attended recommended counseling. The mother attended a parenting class with Jeffrey and Benjamin. According to the instructor, the children enjoyed being with their mother and protested when it was time to leave. The mother received the highest grade on the written test and thoroughly understood the techniques taught which were primarily based upon behavior modification. The instructor said that mother was very eager to learn how to help Jeffrey.
One-stop Child Advocacy Centers (CAC) are designed to help alleviate many of the inherent conflicts in the current child protection system. Child Advocacy Centers’ number one goal is to reduce trauma to the child abuse victim by coordinating a child’s interview to include professionals from multiple agencies, which can reduce the number of interviews and improve the quality of the investigation. They help children avoid the trauma of repeating their story at various stops along the legal and judicial path.
. Mich. Comp. Laws § 722.628(6) (2008) (mandating that child protective services investigators use model interview protocol); State of Mich. Governor’s Task Force on Children’s Justice & Dep’t of Human Servs., Forensic Interviewing Protocol (2005), available at https://www.michigan.gov/DHS/0,1607,7-124-5452_7119_25045—,00.html.
. Vivek Sankaran, Wells Conference on Adoption Law: Judicial Oversight Over the Interstate Placement of Foster Children, 38 Cap. U. L. Rev. 385, 391 (2009) (“the unfortunate reality is that child welfare procedures are often conducted by young, inexperienced workers who lack specialized training and carry high caseloads”).
. The Oregon Interviewing Guidelines (OIG) were originally developed by professionals at the request of the Health Advisory Council on Child Abuse, a group convened by the state legislature to ensure that child abuse evaluators in Oregon were highly skilled and well-trained. The OIG was published in 1998 for a target audience of assessment center-based interviewers. The 2004 revision expanded the document to address all professionals—including both law enforcement officers and child welfare workers charged with conducting field, investigations.
. Or. Dep’t of Justice, Oregon Interviewing Guidelines 3 (2012), available at http://www.doj.state.or.us/victims/pdf/oregon_interviewing_guidelines.pdf.
. The flexibility noted by the Oregon protocol suggests regional differences in child abuse investigations. However, it is unclear from the protocol, or other research, why the format of the interview would need to be different based on location. While individual interviews may need to be tailored to the child based on many factors including, age, disability, and language, there is no reason why the structured format of interviews in, say, Ashland, Oregon would need to be substantially different from those in, Baker City.
. Mich. Comp. Laws § 722.628(6) (2005) (mandating that child protective services investigators use a model interview protocol); State of Mich. Governor’s Task Force on Children’s Justice & Dep’t of Human Servs., Forensic Interviewing Protocol (2005), available at https://www.michigan.gov/DHS/0,1607,7-124-5452_7119_25045—,00.html. It is unclear how Michigan treats initial interviews with a child who is the subject of a child welfare investigation, or whether even these contacts must follow the interview protocol.
. Mich. Exec. Order No. 2010–18, available at http://www.michigan.gov/granholm/
0,4587,7-168-21975-242311–,00.html; Governor’s Task Force on Child Abuse and Neglect, Michigan.gov (2010) http://www.michigan.gov/dhs/0,4562,7-124-7119_50648_66367—,00.html.
. Birckhead, supra note 90, at 417. See also Frank E. Vandervort, Videotaping Investigative Interviews of Children in Cases of Child Sexual Abuse: One Community’s Approach, 96 J. Crim. L. & Criminology 1353, 1415 (2006) (stating that “[videotaping] serves the interests of the community, as it achieves a fair and just result for victims, suspects, and defendants”).
. Michigan does not tie admissibility to the forensic protocol—that is, there is no requirement that protocol is followed in order for a statement to be deemed admissible. Mich. Comp. Laws § 3.972(C)(2)(a) (1985); In re Britton, 306495, 2012 WL 1648836 (Mich. Ct. App. 2012); In re D.K., 289371, 2009 WL 3401152 (Mich. Ct. App. Oct. 22, 2009); In re Archer, 744 N.W.2d 1, 8 (Mich. Ct. App. 2007).
. In re Stephen M., W10CP02014057A, 2008 WL 5572975 (Conn. Super. Ct. Dec. 31, 2008) (“Many of the horrendous problems in the present case would have been obviated by a professional interview recorded on video tape.”); In re Interest of W.S., 899 S.W.2d 772, 778 (Tex. App. 1995) (“Based on our review of the videotape, the trial court properly admitted the videotape, and therefore, the trial court did not abuse its discretion by admitting the tape into evidence.”).
. Commonwealth v. Delbridge, 855 A.2d 27, 39–41 (Pa. 2003) (holding that hold that taint is legitimate question for examination in cases involving complaints of sexual abuse made by young children and finding pretrial exploration of taint is necessary when there is some evidence that improper interview techniques, suggestive questioning, vilification of accused and interviewer bias may have influenced child witness to such degree that proffered testimony may be irreparably compromised); Matter of Dependency of A.E.P., 956 P.2d 297 (Wash. 1998); English v. State, 982 P.2d 139, 146–47 (Wyo. 1999) (declining to adopt separate pretrial taint procedure, but holding that existing statutory competency hearing would allow evidence of taint).
. See, e.g., In re R.M., 718 N.E.2d 550, 556 (1999) (“There is no evidence in the record to suggest Ellen was coached or coaxed into making the statements describing the physical abuse. We find Ellen’s statements were sufficiently reliable to support a finding of physical abuse.”).
. See, e.g., People v. Montoya, 57 Cal. Rptr. 3d 770 (2007) (“The relevant determination in California is whether a minor victim is competent to testify. . . . The capacity to perceive and recollect is a condition for the admissibility of a witness’s testimony on a certain matter, rather than a prerequisite for the witness’s competency.”); United States v. Geiss, 30 M.J. 678, 681 (1990) (“[W]e hold that evidence of suggestive questioning or coercive pretrial interviews goes to the credibility of a witness rather than to the admissibility of testimony.”); State v. Olah, 767 N.E.2d 755, 758 (Ohio Ct. App. 2001) (“No Ohio appellate court has either followed Michaels or independently determined that a pretrial taint hearing is required if a child witness is potentially contaminated.”); State v. Bumgarner, 184 P.3d 1143, 1150–51 (Or. Ct. App. 2008) (“Whether a person, who has the ability to perceive an event, recall it and relate the recollection will tell the truth is to be tested by cross-examination and not by a motion to disqualify the witness as incompetent. The competency inquiry should be made with a view to the preference toward allowing the trier of fact to be the ultimate judge of the quality of the evidence.”).
. See, e.g., Annette R. Appell, Representing Children Representing What?: Critical Reflections on Lawyering for Children, 39 Colum. Hum. Rts. L. Rev. 573, 575 (2008) (“More recently, the notions that children have individual voices and that their voices should be heard have gained currency through the expressive power of the law, most universally in the Convention on the Rights of the Child.”); Linda D. Elrod, Client-Directed Lawyers for Children: It Is the “Right” Thing to Do, 27 Pace L. Rev. 869, 891 (2007) (arguing that “the child’s voice” should be added to debate “whenever the child’s interests and the parent’s interests are not aligned,” for example in “abuse and neglect situations”).
. Convention on the Rights of the Child, United Nations Human Rights (Nov. 20, 1989), http://www.ohchr.org/EN/ProfessionalInterest/Pages/CRC.aspx
. Convention on the Rights of the Child, G.A. Res. 44/25, U.N. GAOR, 44th Sess., Supp. No. 49, U.N. Doc. A/Res/44/25 (1989), available at https://treaties.un.org/pages/ViewDetails.
. For example, the National Association of Counsel for Children has recommended that “[c]hildren need to be involved in the entire litigation process.” See NACC Recommendations for Representation of Children in Abuse and Neglect Cases; American Bar Association Standards of Practice for Lawyers Who Represent Children in Abuse and Neglect Cases 1, 9 (2003); Appell, supra note 212, at 587.
. Donald Duchette & Julian Darwall, Child Representation in America: Progress Report from the National Quality Improvement Center, Improve Children Rep., http://www.improvechildrep.org/Portals/0/DuquetteFLQSpr12-Final.pdf (forthcoming) (“It is widely accepted that children require attorney representation in dependency proceedings. This consensus is based on the practical necessity of attorneys in negotiating complex judicial proceedings, the state’s interest in or child’s right to empowerment through participation, constitutional arguments or analogy to other legal contexts, and the therapeutic nature of the attorney-client relationship.”). But see Katherine Hunt Federle, The Ethics of Empowerment: Rethinking the Role of Lawyers in Interviewing and Counseling the Child Client, 64 Fordham L. Rev. 1655, 1658 (1996).
. Pew Commission on Children in Foster Care, Fostering the Future: Safety, Permanence, and Well-Being for Children in Foster Care 43 (2004); Emily Buss, You’re My What? The Problem of Children’s Misperceptions of Their Lawyers’ Roles, 64 Fordham L. Rev. 1699, 1746 (1996); Davin Youngclarke, Kathleen Dyer Ramos & Lorraine Granger-Merkle, A Systematic Review of the Impact of Court Appointed Special Advocates, 5 J. Center for Fams., Child. & Cts. 109 (2004).
. Quality Improvement Ctr. on the Representation of Children in the Child Welfare Sys., Needs Assessment Literature Review 5 (2010), available at http://www.improvechildrep.org/Portals/0/QIC%20Child%20Rep%20Literature%20Review.pdf.
. Santosky v. Kramer, 455 U.S. 745, 768 (1982) (“We cannot believe that it would burden the State unduly to require that its factfinders have the same factual certainty when terminating the parent-child relationship as they must have to suspend a driver’s license.”). Yet there is no similar constitutional requirement that states prove the original case of abuse or neglect by clear and convincing evidence.
. In re Cindy L., 947 P.2d 1340, 1348–49 (Cal. 1997) (noting several difficulties with proving child sexual abuse and finding that, because of these difficulties, “categorical exclusion of child hearsay, or admission only if the hearsay fits within traditional yet narrow categories” will result in the exclusion of evidence necessary to prevent abuse); Ann Cederborg, Delay of Disclosure, Minimization, and Denial of Abuse When Evidence is Unambiguous, in Child Sexual Abuse: Disclosure, Delay, and Denial 170 (Margaret-Ellen Pipe, Michael E. Lamb, Yael Orbach, & Ann-Christin Cederborg, eds., 2007).
. Cederborg , supra note 320, at 170 (noting that instructions to keep secret have powerful effect on five- and six-year old children and that perpetrators can use their position of power over children to conceal abuse).
. Lyon & Saywitz, supra note 4, at 835 (“Child victims may be disproportionally likely to be called to testify in sexual abuse cases because corroborative evidence more often exists for other forms of maltreatment, such as physical abuse and neglect, or other crimes, such as domestic violence and homicide.”).
. Jennifer Anderson, Julie Ellefson, Jodi Lashley, Anne L. Miller, Sara Olinger, Amy Russell, Julie Stauffer, & Judy Weigman, Forensic Interview Protocol: RATAC, 12 T.M. Cooley J. Prac. & Clinical L. 193, 193 (2009–2010). See also A.P. v. Dep’t of Pub. Welfare, 696 A.2d 912, 915–16 (Pa. Commw. Ct. 1997) (finding hearsay statements of child admissible while noting that, “[i]t must be recognized at the outset that cases such as this present special concerns, none more important than the welfare of the children involved”).
. Child Welfare Information Gateway, Addressing Racial Disproportionality in Child Welfare (2011), available at https://www.childwelfare.gov/pubs/issue_briefs/racial_
. Richard Wexler, Take the Children and Run: Tales from the Age of ASFA, 26 New Eng. L. Rev. 129, 137 (2001). The 2011 Child Maltreatment study measured “Percentage of Foster Care Children Who Were Not Victimized by a Foster Care Provider.” Child Maltreatment 2011, supra note 2, at tbl. 3–15. Most children in foster care were not victimized, but some states had a greater than one percent rate of victimization in 2011. Id.
. Emily Buss, Parents’ Rights and Parents Wronged, 57 Ohio St. L.J. 431, 439 (1996). See also Trudy B. Festinger, The 24-Month Court Review of Children in Foster Care, 54 Child Welfare 211 (1975).