A Case Concerning Children's False Memories of Abuse: Recommendations Regarding Expert Witness Work

Abstract

Expert witnesses can play a major role in legal cases concerning the reliability of statements. Abuse cases frequently contain only the memories of eyewitnesses/victims without the presence of physical evidence. Here, it is of the utmost importance that expert witnesses use scientific evidence for their expert opinion. In this case report, a case is described in which 20 children reported being sexually abused by the same teachers at their elementary school. The investigative steps that were taken by the police and school authorities are reviewed, including how they probably affected memory. In order to provide a sound expert opinion regarding the reliability of these statements, three recommendations are proposed. To reduce the effect of confirmation bias and increase objectivity, it is argued that expert witnesses’ reports should contain alternative scenarios, be checked by another expert, and focus on the origin and context of the first statement.

Legal proceedings in which children testify that they have been sexually victimized are unique cases. These are frequently cases in which physical evidence (e.g., DNA) is absent and children's statements are diametrically opposed to the statements of the suspect (Brackmann, Otgaar, Sauerland, & Jelicic, 2016; Howe, 2013). In such proceedings, it is of vital importance to assess the reliability of children's statements. This is because the lynchpin in these cases is whether children's statements reflect an authentic experience or whether such accounts are tainted by suggestive influences and memory errors (i.e., false memories).

Among such child sexual abuse proceedings, there is a select group of cases in which a large number of children report having been maltreated by the same person(s) at an elementary school or daycare center (Garven, Wood, Malpass, & Shaw, 1998). During the 1980s, an ‘outbreak’ of such ‘daycare abuse cases’ occurred across the United States and Europe (Bottoms & Davis, 1997). The bottom line in many of these cases was that children often made false accusations based on false memories that arose during suggestive interviewing (for a recent review, see Howe & Knott, 2015). Oftentimes, such cases were characterized by bizarre sexual abuse allegations supposedly committed by a single perpetrator. In the last few years, this issue of these special abuse cases has gained renewed attention. Specifically, in many Western countries, similar cases have surfaced in which public figures were publicly accused of having abused many children in previous decades (e.g., the Dutch Roman Catholic Church case; Saunders & Fivush, 2015). The central question in these cases is to estimate the reliability of statements of alleged (child) victims.

In the present case report, a recent situation which has similar contours to these earlier child abuse proceedings is described. The research literature on children's memories for abuse has greatly expanded since the 1980s, so in theory an expert witness today should be able to provide a report with a much better empirical foundation compared to the experts of the 1980s. But applying research evidence to an individual legal case is less straightforward than it seems. The foremost aim of this case report is to arrive at some specific recommendations on how to deal with children's statements of abuse when (legal) psychologists are asked to act as expert witnesses. The paper is organized as follows. First, an explanation of how children's statements can be adversely affected by the formation of false memories is presented. After this, some of the historic child abuse cases that involved false and suggested memories are described. The current case is then presented, along with an explanation of how expert witness work is conducted in different jurisdictions, followed by recommendations for how future proceedings – should they arise – could be better handled.

Eliciting False Memories

A burgeoning literature exists on ways of how to elicit false memories in the psychological laboratory. Of importance, these false memory methods have been invented to mimic real-life circumstances – including, for example, suggestive police interviewing tactics. One of the methods most often used to induce false memories is the misinformation paradigm (Loftus, 2005). In this paradigm, participants receive some stimuli (e.g., a video of a robbery) or are involved in an interactive event (e.g., participating in a magic show). Following this, they receive suggestive information (i.e., misinformation: ‘Was the culprit carrying a gun?’ when no gun was present, or ‘Was the magician wearing a red cape?’ when there was no cape) about the witnessed event. In the final stage, participants’ memories of the original event are tested. The standard finding is that many participants incorporate the misinformation into their memory reports, thereby producing a false memory.

Although the misinformation paradigm elicits false memories for details, the implantation paradigm is known for its elicitation of rich false memories for entire events (e.g., Loftus & Pickrell, 1995; Otgaar, Candel, Merckelbach, & Wade, 2009). In this paradigm, participants are asked what they can remember about a fictitious event (e.g., a hot air balloon ride). Then, during multiple interviews, they have to report everything they can recollect about the false event. Research has shown that such suggestive interviews lead to 30 to 40% of participants falsely remembering that they experienced the false event (e.g., Otgaar, Scoboria, & Smeets, 2013; Wade, Garry, Read, & Lindsay, 2002).

What these false memory paradigms have in common is that they use suggestions about things that did not happen to evoke false memories. Recently, much attention has been directed towards understanding the production of spontaneous false memories that arise without external suggestive pressure (Brainerd, Reyna, & Ceci, 2008). A popular method used to induce spontaneous false memories is the Deese/Roediger-McDermott (DRM) paradigm (Deese, 1959; Roediger & McDermott, 1995). In the DRM paradigm, word lists containing associatively-related concepts (e.g., fear, emotion, hate, fight) are provided to participants. These concepts are connected to a common theme word called the critical lure (i.e., anger). A plethora of research has shown that many participants falsely remember the critical lure during subsequent free recall and recognition tests (Gallo, 2010).

The above-sketched false memory procedures have also been used with children. These studies have revealed that, in many cases, children are exceptionally vulnerable to suggestive pressure, and hence the production of false memories (Bruck & Ceci, 1999; Otgaar, Candel, Smeets, & Merckelbach, 2010). Furthermore, many of these studies have shown that false memories induced by suggestion are more likely to occur in younger children than in older children and adults (Ceci & Bruck, 1993; Otgaar & Candel, 2011; but see Otgaar, Howe, Brackmann, & Smeets, 2016). In contrast to this age-related pattern for suggestibility, developmental studies on spontaneous false memories have demonstrated that adults are more likely to create false memories than children, a phenomenon dubbed ‘developmental reversal’ (Brainerd et al., 2008).

In legal cases, both suggestion-induced and spontaneous false memories can surface in the statements of children and adults. Both children and adults can be suggestively interviewed by parents, therapists, or police officers, and hence such interviews might negatively affect their statement quality. Although research shows that when children (or adults) spontaneously report having been maltreated, their statements often possess a high level of accuracy (e.g., Bidrose & Goodman, 2000; Leander, Christianson, & Granhag, 2007; Orbach & Lamb, 1999), recent work on false memories shows that adults (and to a lesser extent children) can be prone to producing spontaneous memory errors in their reports. Understanding how such false memories arise is crucial for legal professionals, as they often have incorrect ideas about how memory works (Odinot, Boon, & Wolters, 2015; Ost, Easton, Hope, French, & Wright, in press). A specific example of this is that legal professionals sometimes incorrectly assume that children are inferior eyewitnesses because of their extreme proneness to false memories, thereby discrediting statements of young children (for a recent case report on this issue, see Brackmann et al., 2016).

False Memory Cases

Although during the 1980s the number of cases of alleged abuse at daycare centers increased considerably, only a few of them sparked extensive attention among memory researchers, legal professionals, and the media. One prominent case was the McMartin Preschool trial (Garven et al., 1998; Garven, Wood, & Malpass, 2000). In 1983, seven teachers from the McMartin Preschool located in Los Angeles were charged with sexually abusing more than 100 children in the most horrifying manner. Children ‘remembered’ some extremely bizarre experiences, such as seeing dead and burned babies, riding naked on a horse, and flying to an isolated farm where they had to engage in group sex (Schreiber et al., 2006).

The accuracy of the children's reports was highly contested because of several contaminating influences. First, after the first allegation, parents received a letter in which they were asked to question their children to see whether they were victimized. There is research suggesting that parents often ask children suggestive questions when abuse is suspected (Korkman, Juulosa, & Santtila, 2014). Second, children were also interviewed in a highly suggestive manner by social service workers, which might have led them to make false statements. An abundance of research reveals that children are particularly susceptible to suggestive questioning, and hence the formation of false memories (Bruck & Ceci, 1999). Most of the charges against the teachers were eventually dropped and none of the teachers was convicted of a sexual crime.

In another highly publicized daycare abuse case, nursery school teacher Margaret Kelly Michaels was accused of sexually abusing children at the Wee Care Nursery School. Like the McMartin Preschool case, some of the allegations made by the children were quite bizarre. For example, she was accused of licking peanut butter off children's genitals, playing the piano while being nude, making children drink her urine and eat her feces, and raping children with Lego blocks, knives, forks, and spoons. According to the children's accounts, she abused them during school hours over a period of seven months.

As in the McMartin Preschool case, there are several elements in this case that likely infected children's accounts of sexual abuse. First, children's parents received a letter from the daycare center informing them that some children had stated they were abused. Second, during a subsequent meeting, a social worker told parents that sexual abuse of children is very common and that the parents should examine their children for signs of sexual abuse (e.g., nightmares, bed-wetting). After this, the number of allegations against Kelly Michaels increased significantly. Kelly Michaels was sentenced to 47 years’ imprisonment but was released after 5 years because a number of forensic psychologists sent an amicus letter to the court expressing concerns regarding the accuracy of the children's sexual abuse statements (Bruck & Ceci, 1995).

Daycare abuse cases are not confined to the United States. Similar cases have arisen in European countries as well. For example, in 1987, in the small Dutch village of Oude Pekela, children reported being sexually molested by a group of satanic clowns (Goodman et al., 1997; Jonker & Jonker-Bakker, 1991). In this case, the idea was that these accusations were the result of suggestive interviewing tactics. Other notable examples are the Wormser and Montessori trials in Germany during the 1990s, in which children's reports of abuse were likely influenced by suggestive interviewing techniques as well (Schade & Harschneck, 2000). In the present study, a recent child sexual abuse case that has many similarities with previous daycare abuse cases – and in which the authors Henry Otgaar (HO) and Corine de Ruiter (CdR) were involved as expert witnesses – is analyzed.

The ‘Galileo’ Case

In this case,^1. HO and CdR were contacted by parents of alleged child victims of sexual abuse. The parents notified them that their children had been sexually abused at the Galileo Elementary School. This school is located in a small town in the western part of the Netherlands. In September 2009, a father of one of the alleged victims filed an official charge with the police on behalf of his 8-year-old son, Guy. The father declared that his son had been sexually abused by two teachers at Galileo. After the charges had been filed, the police interviewed more children, who also reported having been abused. In total, 20 children reported having been abused by the same two teachers at the school.

When reading the case file, a number of critical facts became evident. In cases such as this, it is essential that experts examine how the first statement of abuse was produced (e.g., see Volbert & Steller, 2014). This information is important because, for children, statements that occur spontaneously are often provided without any (suggestive) intervention and may provide a more accurate account of what was experienced (e.g., Bidrose & Goodman, 2000). Furthermore, it is not uncommon that such spontaneous statements are provided after long delays, as child victims of sexual abuse are often reluctant to talk about their abusive experiences (London, Burck, Ceci, & Shuman, 2005). Interestingly, in the Galileo case, the first statement that was provided by the child had nothing to do with sexual abuse – that is, in 2008, Guy stated that a female teacher named Maria had punished him in class. Guy talked about this to his parents. Specifically, Guy was at times not allowed to go to lunch or to the toilet. Maria confirmed that this had happened, and other teachers also reported that Maria was a strict teacher.

Because Guy had many problems in and out of school (i.e., he was highly anxious about going to school), he had been undergoing therapy. The therapist concluded that Guy had experienced something highly traumatic and that he had post-traumatic stress disorder (PTSD). To uncover the traumatic experiences, the therapist used anatomically-detailed dolls and Guy was asked to show whether he was perhaps (sexually) abused. A large body of research shows that such dolls are highly suggestive, lead to fantasy play behavior, and increase the chance of incorrect reports (e.g., Otgaar, Horselenberg, van Kamper, & Lalleman, 2012; Poole & Bruck, 2012). Hence, to date, the use of such props is discouraged in interviews with child victims of alleged sexual abuse (Bruck, 2009; Poole, Bruck, & Pipe, 2011; see also Otgaar, Van Ansem, Pauw, & Horselenberg, in press). Although Guy did not come up with any details concerning sexual abuse during the therapy session with the anatomically-detailed dolls, it was only after this (suggestive) therapeutic session that Guy started to change his statement about punishment into statements concerning sexual abuse.

One day, after the first official police charge, the school sent a letter to all parents about possible abuse at the school. In this letter (see the Appendix for the text of the letter), the parents were informed of the alleged abuse and told that more children could have been victimized. Moreover, the parents were told that they would receive information on how to interpret signs that their child was abused. After the letter, many other children came forward, having ‘remembered’ being sexually abused at the school. It is striking that in many earlier daycare abuse cases (e.g., the McMartin Preschool, the Wee Care Nursery School), such (suggestive) letters had also been sent to parents prior to their children making similar accusations of having been sexually abused. Also, it is important to remember that the idea that there are specific signs indicative of (sexual) abuse has not received any empirical support. In fact, research shows that there are no reliable signs that are diagnostic of child sexual abuse (Kelly, Koh, & Thompson, 2006; Kendall-Tackett, Williams, & Finkelhor, 1993).

In the present case as in the earlier ones, the children's statements contained certain bizarre details. For example, Guy reported that Maria had cut off a piece of his penis. Furthermore, some children remembered that they had to go to one of the teachers’ houses where they had to look at crocodiles. Among some clinicians, there is the idea that children do not make up such bizarre stories and hence when bizarre details do arise, they must be authentic (Dahlenberg, 1996; Everson, 1997). However, research shows that false memories for highly implausible events can easily be evoked, even in children (Otgaar et al., 2009; see also Mazzoni, Loftus, & Kirsch, 2001).

A final remark pertains to the fact that many of the children's parents talked on many occasions about the alleged sexual abuse. They even held a meeting to discuss their concerns. Numerous studies have shown that such discussions might inadvertently affect children's statements (Principe & Schindewolf, 2012). More precisely, these studies indicate that discussions might lead to rumors concerning, for example, sexual abuse. When children are subjected to such rumors, this can lead them to falsely remember that they too were abused (Principe & Schwindewolf, 2012).

Taken together, this case has many parallels with previous daycare abuse cases. Although Guy did spontaneously report experiences concerning (harsh) punishment, his subsequent statements are likely to have been adversely affected by what is considered to be obvious indications of suggestive influence. These include the distribution of the letter to all parents by the school alleging sexual misconduct and the suggestive methods used by Guy's therapist. At first, the case was not prosecuted because of the lack of convincing proof that these crimes had taken place, and because a legal psychologist wrote a report casting doubt on the reliability of the children's statements. Following this, the parents contacted HO and CdR and went to the Court of Appeal. The Court of Appeal decided that because the expert witnesses concluded that the children's statements could have been affected by suggestive influences, the teachers would not be prosecuted. In cases like the one described here, expert witnesses have a crucial and often decisive role in determining whether the statements are reliable. Hence, expert witnesses should look carefully at all case files in order to provide an objective judgment about the reliability of children's statements.

Expert Witness Work in Different Countries

The ‘Galileo’ case took place in the Netherlands. This is important to point out, because expert witnesses can play somewhat different roles depending on which country they are in and which criminal legal system pertains to that country. Broadly speaking, criminal law systems from across the world can be divided into two types: inquisitorial and adversarial (De Ruiter & Kaser-Boyd, 2015). The Netherlands uses the strictest inquisitorial system in the world, whereas countries like the United Kingdom and the United States adopt an adversarial approach. In the adversarial system, legal trials can be regarded as contests between equivalent parties (Van Koppen, 2007): the prosecution and the defense. They argue their case in a somewhat equal role in front of the judge (and oftentimes, the jury). Under the adversarial model, the judge is considered to have a neutral position, intervenes only in the case of conflicts, and is a purveyor of all things to do with the law.

The adversarial system is different from the inquisitorial one. Here, there is no jury. Also in, for example, the Netherlands, a trial is not about a fair contest between parties but about finding the truth (Van Koppen & Penrod, 2003). The judge is a professional in that, compared to adversarial systems, judges in countries like the Netherlands receive special training. They have to attend law school for five years and then undergo an additional six years of special judicial training. Part of this extra training involves courses concerning the psychology of law (Van Koppen, 2007). Furthermore, in the inquisitorial system, the judge has an active role and leads the investigation. Furthermore, the prosecution aims to come up with an independent judgment before a case can be continued, and this can mean that a prosecutor asks for an acquittal. Also under this model, defense attorneys and prosecutors are not the ones primarily asking questions in the courtroom. Because the judge has an active role, he or she is primarily in charge of asking questions of the defendant.

Expert witnesses have a rather unique position in each of these legal systems. That is, in both systems, expert witnesses can be called upon to appear in court, and in both systems expert witnesses can be approached by either the prosecution or the defense. However, one crucial difference is that in the adversarial system there is the chance that expert witnesses are somewhat biased to the side that retained them (Saks, 2003). Indeed, recent research has shown that such allegiance biases occur in expert witnesses (Murrie, Boccaccini, Guarnera, & Rufino, 2013). Such biases might lead to battles between experts that were hired by opposing parties. Although in the inquisitorial system expert witnesses are assumed to be more independent (Saks, 2003) and are frequently not asked to elaborate on their expert witness report in the courtroom, this does not prevent them from battling with other experts in the courtroom.

To illustrate, a recent case is described in another paper in which HO was asked to evaluate the reliability of a series of statements made by a 6-year-old girl (Brackmann et al., in press). The basic conclusion of this expert report is that there are no strong reasons to doubt the statements of the girl. The defense had hired another expert, a clinical psychologist, who reasoned that the girl's statements were contaminated because of autosuggestion (i.e., similar to the effect discussed earlier in this paper in terms of spontaneous false memories that arise without any external suggestive pressure). However, during the trial, HO made the point that recent research shows that autosuggestive effects are less likely to occur in young children than in older children and adults (Brainerd et al., 2008), and that it is unlikely that autosuggestion has played a role in the statements of the child in this case. Eventually, this resulted in a disagreement between the two experts in the courtroom concerning the susceptibility of false memory production in children and adults.

Of course, part of the reason why there may be disagreement between experts is related to the fact that experts are often experts in different domains (e.g., the science of memory vs. clinical psychology), and issues can arise as to which expert testimony is admitted in court. For example, in the United States, only expert evidence which meets the Frye or Daubert criteria is allowed (e.g., falsifiability, known error rate, subjected to peer review and publication; Saks & Faigman, 2005). Although such specific criteria are absent in the Netherlands, expert databases are currently being constructed which contain expertise in various fields (e.g., forensic psychology) and which have been demonstrated to be accepted by the scientific community. Nonetheless, in both systems, it remains possible that expert witnesses are commissioned who lack the scientific knowledge about the subject they are asked to comment on.

Before being asked to review the Galileo case, another report was prepared by a legal psychologist. The parents argued that the first psychologist had not looked at all of the necessary case files in order to provide an appropriate and objective conclusion. Hence, after the parents contacted the authors of the present study, it was decided not to look at the earlier report in order to prevent the development of any expectancy effects about the case, and thus to look at all the case files independently. Finally, when a comparison was made between the expert witness report of the present authors and the previous report, many similar arguments were found, along with a highly similar conclusion that there are elements in the case that could have affected the children's reports.

Based on this case and other expert witness work in which the present authors have been involved, it is believed important to strengthen expert witness decisions by offering a number of recommendations. It is thus proposed that the following recommendations need to be implemented by expert witnesses in general, whether they are working in adversarial or inquisitorial systems.

Recommendations

In the majority of child sexual abuse cases, the central question revolves around whether statements are reliable or not. In order to provide a well-balanced conclusion concerning the reliability of statements, three recommendations are proposed. It is suggested that when adopting these recommendations, expert witnesses are in a better position to provide an unbiased conclusion based on the latest scientific findings.

The first recommendation is that an expert witness should commence with so-called ‘scenario-building’ (Crombag & Wagenaar, 2000; Raskin & Esplin, 1991; Rassin, 2014). It is proposed that when starting with expert witness work concerning the reliability of testimony, one should postulate at least two scenarios. One scenario would signify the ‘guilty’ scenario, in which the assumption is that statements refer to an authentic experience.^2. The alternative scenario could be called the ‘innocent’ scenario, in which the assumption is that statements are the result of fabrication (e.g., false memories, deception). Obviously, such a recommendation is not only relevant for expert witnesses; more generally, it should be used throughout the entire investigative process and is recommended for the police during the interview process as well.

The idea behind this is that when expert witnesses are working on a case concerning the reliability of statements, they should decide whether they find elements in the case supporting the ‘guilty’ or ‘innocent’ scenario. If they find more support for the ‘innocent’ scenario then there are reasons to doubt the reliability of the statements. However, if more support is found for the ‘guilty’ scenario then there are few reasons to doubt the reliability of statements. To give an example, consider the scenario where the expert witness is reading a case file concerning the reliability of a suspect's confession and discovers that the suspect only confessed to the crime after a suggestive interrogation by the police. Because research shows that this technique might lead to a false confession (Kassin, 2008), the expert witness has found a piece of evidence that supports the alternative ‘innocent’ scenario.

Why the construction of such scenarios is important is because it might protect the expert witness from expectancy effects or confirmation bias (Kassin, Dror, & Kukucka, 2013). Specifically, by using such scenarios, the expert witness is not solely focused on, for example, finding evidence for proof of false memories, or the reverse. Also, it might guard expert witnesses from the allegiance bias in which forensic experts are unconsciously biased towards the party that they are retained by (i.e., prosecutor vs. defense; Murrie et al., 2013). Indeed, although not in the realm of expert witness work, there is evidence suggesting that the postulation of alternative hypotheses immunizes against confirmation bias. Specifically, O'Brien (2009, Study 2) provided participants with a police file of a criminal investigation, and their task was to answer certain questions about the case. Participants were assigned to different conditions, but of relevance here is the condition in which participants were asked who committed the crime (generating a hypothesis) and the condition in which participants also had to think about why the person might be innocent (alternative hypothesis). O'Brien found that a statistically significant greater amount of bias was observed in the generate-a-hypothesis group compared to the alternative-hypothesis group, inasmuch as participants in the generate-a-hypothesis group remembered more facts consistent with the guilt of the suspect than the alternative-hypothesis group.

Recent research has also shown that thinking about alternatives might make jurors more critical about eyewitness evidence. That is, Rodriguez and Berry (in press) provided mock jurors with a brief crime summary containing high-quality (e.g., unbiased line-up instructions) or low-quality (e.g., biased line-up instructions) eyewitness evidence. Then, half of the participants had to adopt a counterfactual thinking style (i.e., considering how the present situation could have been different) which is closely linked to coming up with an alternative scenario. The other half did not have to engage in counterfactual thinking. The most important result is that when the participants received low-quality eyewitness evidence, they were least likely to reach guilty verdicts when they used a counterfactual mindset. Collectively, these studies suggest that alternative scenarios might lead people to search for different outcomes, thereby potentially reducing confirmation bias.

There have been attempts to increase the relevance of scenario-building (Volbert & Steller, 2014). Although these attempts are important, they have been rather unsuccessful. This is because expert witnesses are often unclear as to exactly when this approach should be used and how many scenarios should be constructed (Neal & Brodsky, 2016). It is suggested that scenario-building should be undertaken before reading the case files – so, in its simplest form, only two scenarios should be proposed at the start of the case: one referring to statements based on a true experience and one indicative of a false statement. Of course, the term ‘false’ can suggest different issues, such as intentional fabrication (i.e., lies) or unintentional fabrication (i.e., false memories). However, this nuance is not relevant at the beginning, as expert witnesses commonly have no prior knowledge concerning the case. After reading the entire case file, additional (exploratory) scenarios can be generated that provide more nuances in scenario building. To give a concrete example, suppose that Expert Witness X is asked to provide his or her expert opinion on statements of several eyewitnesses to a crime (e.g., theft). Expert X postulates two scenarios (guilty vs. innocent). When reading the case files, Expert X discovers that the eyewitnesses had multiple arguments with the suspect before the theft, and that they were not actually present during the crime. Also, Expert X finds out that the eyewitnesses have had many meetings with each other to discuss the case. These particular elements might lead to an additional (exploratory) scenario referring to idea that the eyewitness statements are not only false but have been affected by deception as well.

The idea of postulating scenarios before reading the case files comes close to the concept of preregistration in science (Wagenmakers, Wetzels, Borsboom, van der Maas, & Kievit, 2012). Because of the increased attention on failed replications in psychology, the idea of preregistration is gaining momentum in the psychological field. The idea behind preregistration is quite simple: researchers indicate before conducting an experiment the hypotheses and planned analyses at a public forum. After data collection, exploratory analyses can be conducted but the preregistered hypotheses (and analyses) are the most relevant for the experiment and fall under the label of confirmatory research. To some extent, this is exactly what is done during scenario-building in expert witness work. Admittedly, it is true that no empirical knowledge yet exists on whether the use of this preregistration strategy results in better expert witness work. However, it would be useful to test such hypotheses in future research.

The model of ‘scenario-building’ is also highly linked with hypothesis-testing in many scientific disciplines. That is, before conducting an experiment, it is advisable to formulate a null hypothesis and an alternative hypothesis (Gigerenzer, 2004; Schneider, 2015). The null hypothesis is analogous to the ‘guilty’ scenario and the alternative hypothesis refers to the ‘innocent’ scenario. Here too, when enough evidence exists for the alternative hypothesis (e.g., p < .05), the null hypothesis is rejected. Scenario-building also bears many similarities to Bayesian hypothesis-testing, in which one can calculate average likelihoods known as Bayes Factors, which refer to the degree to which the data are more likely under the null than under the alternative hypothesis (Wagenmakers et al., in press). To some extent, this is exactly what scenario-building entails. When using scenario-building, the task of the expert witness is to judge whether a given case file is more likely to contain elements for the ‘guilty’ or ‘innocent’ scenario.

In the case outlined here, it is evident that there is support for the ‘guilty’ scenario in that Guy spontaneously talked about being harshly punished, which was confirmed by his teacher and her colleagues. However, the case also contains a number of elements – e.g., the suggestive letter, meeting with parents, suggestive therapy – that weigh in favor of the ‘innocent’ scenario. Taken together, it was concluded that besides the spontaneous statement of Guy regarding harsh punishment, there are further reasons to doubt the reliability of the statements concerning sexual abuse.

The second recommendation is that after setting up different scenarios, the expert witness should focus on the circumstances around the first statement. This is critical information, as studies have revealed that spontaneous statements often contain a high degree of accuracy (Orbach & Lamb, 1999). Furthermore, when there is information in the case file that the first statement arose after suggestive interviewing, this is also essential information for estimating the reliability of the statement. This is especially true when a case consists of repeated interviews. Repeated interviews by themselves are not necessarily dangerous for the reliability of eyewitness memory (Goodman & Quas, 2008). However, when the first statement is developed after suggestive pressure, follow-up statements are likely to contain erroneous details even if these statements are the result of well-conducted interviews (Ceci, Kulkofsky, Klemfuss, Sweeney, & Bruck, 2007).

A third recommendation for expert witness work is that the draft expert witness report should be read by another blind and unbiased expert. Ideally, such an expert witness should be someone who is also an expert in the field of, for example, eyewitness memory, and who has no knowledge concerning the specifics of the case. The task of this person is to review the report and provide critical comments where appropriate. This is an additional strategy for minimizing the contaminating influence of confirmation bias. Therefore, it is important to note that this second expert is not merely a proof reader; on the contrary, the task of this second expert is to critically check whether the expert witness has carefully used the scientific literature and the information in the case file to support his or her scenarios and conclusion. One might think that having an extra expert review the report amounts to too much work and is therefore not feasible. However, this approach is common practice among Dutch legal psychologists. Another reason why the expert report should be read by a second expert is because the act of scenario-building can lead to some subjective judgments by the expert witness. For example, it is not uncommon for an expert witness to find some support for the ‘innocent’ (e.g., suggestive interviewing) and ‘guilty’ (e.g., first statement was spontaneous) scenarios, at which point the question is how the expert witness should weigh these findings. A second expert can test the manner in which such judgments have been made and examine whether the decision concerning the weighing has any face value.

It is suggested that when using these three rather straightforward recommendations, expert witnesses guard themselves against making faulty inferences and unintentionally focusing more on one specific scenario than the others. The downside to conclusions based on scenario-building is that they are mostly qualitative (i.e., ‘there are no reasons to doubt the reliability of statements’). This results in conclusions being somewhat superficial, as there are no clear cut-off scores for what constitutes an unreliable statement. Of course, there are some exceptions to this disadvantage as, for instance, levels of suggestibility can be quantitatively assessed using the Gudjonsson Suggestibility Scale (GSS; Gudjonsson, 1984) and high levels of suggestibility might be seen as somewhat supportive of the ‘innocent’ scenario. Nonetheless, the method of scenario-building – which is widely used among Dutch legal psychologists (e.g., Rassin, 2014) – could benefit from a more quantitative approach. Rassin (2014) suggests scoring each element (e.g., suggestive interviews) in favor of the ‘innocent’ scenario with one point. The more points a case contains, the more likely the statements are contaminated. Although others have suggested similar strategies (O'Donehue, Benuto, & Cirlugea, 2014), the problem with such an approach is how much weight should be given to each element. For example, if an eyewitness reports having seen the perpetrator of a robbery at night, it is difficult to decide how much weight should be given to the bad lighting conditions, especially if there are no other reasons to doubt the statements. Or consider the following: two adult sisters claim to have been sexually abused by their father when they were 6 years old. One of the sisters came forward with the statement one year after the alleged event, while the other sister took eight years to come forward. For both sisters, their first statement occurred after a delay, but it is quite difficult to decide when a delay can be seen as a reason to doubt the statement or can be regarded as typical for child sexual abuse cases (i.e., delayed disclosure; Goodman-Brown, Edelstein, Goodman, Jones, & Gordon, 2003). Future research is needed to examine whether such a quantitative approach has scientific merit.

Obviously, the recommendations that put forward in this paper are specifically focused on expert witness work. However, other parties could benefit from such recommendations as well. The model of scenario-building is relevant for everyone interviewing alleged victims of sexual abuse (e.g., police, therapists). That is, when taking into consideration that statements might also be false, statements are not routinely accepted as true and are examined more critically. This might reduce the likelihood that suggestive influences occur in child sexual abuse cases. For example, it might reduce the tendency of schools to distribute (suggestive) letters to parents. In another recent abuse case in which HO worked as an expert witness, the director of the elementary school knew about the dangers of suggestive influences on memory. In that case, the school also sent a letter, but no mention was made about the possibility of other children being victimized. Also, correct advice was provided on how children should be interviewed.

To recap, a recent elementary school abuse case has been described which has many similarities with the classic daycare abuse cases from the 1980s and 1990s. From this, it is argued that expert witnesses play a pivotal role in advising courts (whether the system is inquisitorial or adversarial) and that to increase their objectivity and transparency, some simple and easily applicable recommendations should be used during the course of the preparation of the expert report. By embracing these recommendations, not only will such reports be more objective and standardized, but this approach will also enhance the scientific status of psychologists in the courtroom.

Appendix 1. Translated and Edited Version of the Letter Sent to the Parents by the School

Dear parents,

On behalf of the Galileo school, we want to invite you for a special meeting tomorrow evening. We want to inform you about an official charge that a parent made to the police against two staff members of our school. In this letter, we want to inform you about what happened, what the status is of the police investigation, and how we can help you.

What Happened

Last Monday, we were confronted by a parent who stated that two staff members had inappropriate physical contact with his child. The parent also filed a charge at the police about this, and the police have now started an investigation concerning the charge.

Information for Parents

We are deeply shocked about this event. We have to wait for the outcome of the investigation but the accusations are of such a nature that we want to inform you. We want to do this because other children are possibly involved in this case as well.

Support

We are aware that this letter might lead to concern and questions regarding your child. During the meeting, we will inform you about the current status of this case. Professionals will explain how to interpret signs and signals of your child and how you can help your child. Furthermore, advice will be given about what not to do.

Disclosure Statement

No potential conflict of interest was reported by the authors.