How Often Do Children Lie About Being Sexually Abused?

Abstract

The question of the child’s veracity regarding sexual abuse can arise in several contexts. Parents can wonder about their children’s allegations or lack thereof (e.g., given other children’s allegations or worrying signs such as genital rashes). Law enforcement facing decisions about arrest and prosecution also can want to understand this question. Forensic interviewers can be concerned about either about the child’s allegations or lack of allegations. Child protective services facing decisions about removal of children and the safety of children can also wonder about the answer to this question.

The question of the child ’s veracity regarding sexual abuse can arise in several contexts. Parents can wonder about their children’s allegations or lack thereof (e.g., given other children’s allegations or worrying signs such as genital rashes). Law enforcement facing decisions about arrest and prosecution also may want to understand this question. Forensic interviewers can be concerned about either about the child’s allegations or lack of allegations. Child protective services facing decisions about removal of children and the safety of children can also wonder about the answer to this question.

These key individuals may also want to understand the base rate of lying about sexual abuse to make an informed decision about the likelihood or truth telling in an individual case. For example, the reasoning can be, “If only 1 % of children who make a claim of sexual abuse are lying, and this child is alleging abuse then, we ought to proceed with prosecution.” On the other hand, if the base rate is much higher, say, 40 %, then a more cautious approach would be warranted. And the converse is also important, “If x% of children denying abuse are lying when they indicate that they have not been abused but it is actually the case that they have been abused, then perhaps further (maybe even repetitive) questioning and investigation is still warranted.”

We must also recognize two other situations. First, sometimes children allege logically inconsistent states of affairs: at one time they say they have not been abused and at another time they say they were. Because of the logical law of the excluded middle we know that both cannot be true and thus there is interest finding which is true. Finally, a sexual abuse allegation is actually a series of claims. Claims typically involve who abused them, what the abuse consisted of, how many times each kind of abuse occurred, who witnessed this, if anyone, where the abuse occurred (which can be important regarding jurisdiction), etc. We must also countenance that some of these claims can be truthfully put forward while others could be lies .

Further complicating this issue is that some have taken positions on this issue. For example, during the infamous McMartin preschool trial an advocacy organization, “Believe the Children,” was formed by parents involved. One of the central claims of this organization was that “children never lie” about being sexually abused (De Young, 2004). It seems to be the case that their actual position was more (problematically) nuanced, in that they countenanced the possibility that children’s denials may be lies , and this belief was used to justify aggressive, repetitive, and leading forensic interview ing with the children (Schreiber et al., 2006). Thus, they promoted the idea that children’s allegations should be believed without question; while their denials could be lies. In addition, classical psychoanalytic theory (Freud, 1900/1991) suggests that all children go through an Oedipinal stage of psychosexual development in which they want to have sex with their opposite sex parent and may at times confuse fantasy with fact (although as we shall see this may be more indicative of false memories rather than intentionally lying ).

A few key distinctions need to be made at the outset. First, children can lie in either making a allegation (e.g., “My father touched my privates”) or lie in denying this (“My father did not touch my privates”). It is much more common in the literature to recognize and be concerned the latter possibility than the former (see e.g., Summit, 1983 but also see O’Donohue & Benuto, 2012). Second, it is important to point out that lying is only one pathway for a false allegation. Some naively think something along the lines that “if the child utters a claim that they have been sexually abused that is false, then he or she must be lying.” However, this inference is false. What the field has shown over the past three decades or so is that a variety of suggestive influences (e.g., leading questions, repetitive questions, social conformity press, etc) can cause the child to have false memories (e.g., Poole & Lindsay, 1995; Ceci & Bruck, 1993; Quas et al., 2007, see Chaps. 5 and 8 in this volume.). That is, the child “remembers” that x happened to him or her, when in fact, x did not occur. However, it is important to point out that in this case the child is not lying, i.e., intentionally and knowingly stating a falsehood but rather has made a memory error —an error of commission, rather than the more commonly recognized error of omission —forgetting.

Thus, more formally, a key distinction needs to be made. When the question is asked, “How often do children lie about being sexually abused?” we assert the following two criteria need to be met:

• 1. The child is stating a falsehood.

• 2. The child is knowingly stating this falsehood.

This is in direct contrast to a false memory that instead would meet these criteria:

• 1. The child is stating a falsehood.

• 2. The child believes (albeit incorrectly) that what he or she is stating is true.

This second situation is consistent with the false memory research (e.g., Steffens & Mecklenbräuker, 2007, Chaps. 5 and 8 this volume). It is again important to point out that not all false allegations made by children are lies .

As previously stated, a final distinction needs to be made regarding the scope of the lie. Allegations of sexual abuse usually involve many subsidiary claims, e.g., who did it; exactly what did they do; when did they do it;, where did they do it; how many times did they do it; who witnessed it; did they offer any threats or bribes?, etc. The child may lie (or have a false memory ) about any or all of these dimensions. This adds a complexity to this question as in an individual case the child’s allegations may involve a combination of true statements, lies , and/or false memories.

Thus, we are now in a position to see some of the key complexities that need to be addressed before reviewing what is known regarding the question of children’s lying about sexual abuse . Research that attempts to provide information relevant to this question must include:

• 1. Use of a methodology to reasonably conclude that the child ’s claims are, in fact, false. That is, if the child claims that an uncle touched her on the chest in August of 2007; that there is a valid method for determining whether or not this in fact did not occur. As there is no perfect lens into history, researchers need to argue on pragmatic grounds that sufficient sound information was gathered to make a reasonably accurate inference regarding this historical matter.

• 2. Use of a methodology for determining the child ’s state of mind at the time of the claim, namely that the child made a certain claim and knew when he or she was making this claim that it was false. That is, researchers need to distinguish a lie from a false memory as discussed above; and in order to do this, they need to establish with some reasonable amount of evidence that the child was knowingly stating a falsehood.

• 3. Ideally, these two criteria would need to apply to each subclaim in the child ’s allegation. That is, researchers need to countenance the possibility that the child’s claim that their uncle did this was true, but they may be lying about the number of times the uncle abused them.

Meeting these three criteria is not an easy matter and to date we shall argue that extant research has not done a particularly good job in dealing with these.

We also want to briefly address possible motivations for why a child may lie about being sexually abused:

• 1. The child may be seeking to hurt someone by the allegation. For example, they may be angry at this person for a parenting decision.

• 2. The child may be afraid to make the accusation because the perpetrator has threatened them or fail to make the accusation because the perpetrator has bribed them.

• 3. The child may know their accusation could disrupt family life (e.g., financially) and thus falsely deny actual abuse .

• 4. The child may be influenced by an adult who has a stake in the child’s accusation, e.g., a vengeful mother in a custody dispute; or a mother worried about the financial consequences if a perpetrator is incarcerated.

• 5. The child may have a history of lying and this is just one more example of a chronic problem in telling the truth .

• 6. The child may suffer from a mental disorder such as Conduct Disorder in which they disregard societal rules and they do not have normal internalized morality.

• 7. The child may like the attention gained from the false allegation.

• 8. The child may like the secondary gain received from the allegation, e.g., different living situation, etc.

We will now critically review research that has attempted to address this question. In order to find all relevant studies, we used several EBSCOhost databases including PsychINFO and E-Journals, with the search terms “false” + “allegations + sexual + abuse ,” “false + denials + sexual + abuse” “lie + sexual abuse,” and “recantations + sexual + abuse.” This search yielded few relevant results. In fact, we found most of the articles we chose to review in the reference sections of the articles found through EBSCOhost databases. In selecting studies to review, we picked those that utilized participants ages 18 and under and that reported rates of false allegations , false denials and/or recantations in child sexual abuse allegations. No studies specifically looked at rates of false denials; therefore, we chose to include studies that reported rates of denials in validated cases of child sexual abuse, and we referred to those as rates of false denials given that the sexual abuse was considered to have occurred and the child was denying that the abuse took place. We will now examine the criteria used to determine factual accuracy as well as intentionality.

False Allegations

Studies Involving Surveys of Professional Judgments

Kendall-Tackett and New Hampshire Univ. (1991) surveyed 74 law enforcement professionals and 127 mental health professionals to determine what percentage of child sexual abuse allegations were false. The 201 Boston area participants took part in a standardized telephone interview in which they were ask to provide the percentages of children below the age of 6, 6–9, and 10–12 whom they believed to have made false accusations about being sexually abused. Results suggested that most professionals suspected that lying about sexual abuse occurred in less than 5 % of cases, and that children ages 10–12 made more false accusations than their younger counterparts. Eighty-two percent of professionals endorsed that more than 5 % of 10–12 year-olds had lied versus 71 % of professionals for 6–9-year-olds and 59 % professionals for 6-year-olds. Additionally, findings indicated that female professionals reported significantly fewer fictitious allegations than their male peers.

This study has several important limitations and results should be interpreted with caution. First, lying was defined as a child stating that the abuse occurred when in fact it did not (i.e., false positives). Thus, this study did not assess the child’s intentionality and thus failed to distinguish lying from false memory . In addition, the definition used in the study excluded any instances of abuse where the child claimed that the abuse didn’t take place when it clearly did (i.e., false negatives). Second, the only method for determining the percentage of children that had lied about being abused was to ask law enforcement and mental health workers to offer their opinions. No information was provided on how these judgments were made, the correctness of these judgments, or whether the professionals distinguished between telling a falsehood that the child knew to be fallacious and telling a falsehood that the child believed to be true (e.g., false memories). In addition, there is no evidence that a professional’s opinion about rates of lying is a valid indicator of actual rates of lying. Instead, it may be a better index of professionals’ preconceptions about this issue. Additionally, no evidence was presented providing support for the specific reasons the professionals came to their judgments in individual cases, for example that the child was involved in a custody battle where one of the parents had a stake in the allegation. It would also have been useful for professionals to rate the same cases to at least determine interrater agreement on this, i.e., whether multiple professionals agreed that lying did or did not occur in a specific case.

Everson and Boat (1989) interviewed 100 Child Protective Service workers in the Department of Social Services in North Carolina to estimate the rate of false allegations of sexual abuse in CPS cases. In phase 1 of the study, the participants were required to provide estimates of total number child sexual abuse cases in which the CPS worker had participated, number of substantiated cases, and number of cases determined to contain false allegations of child sexual abuse. Eighty-eight of 100 CPS workers returned completed questionnaires. Results indicated that CPS workers reported a total of 1249 cases of child sexual abuse. The mean rate of substantiation, defined as the percentage of cases of child sexual abuse confirmed to be reliable by a CPS investigation , was 56 % across four different age groups (children under 3, between 3 and 6, elementary school aged children, and adolescents). Children were determined to have made false allegations in an average of 4.7 % of the cases. However, there were large discrepancies across age groups, as children under 3 were believed to have lied about being sexually abused in 1.6 % of cases while those older than 12 were thought to have lied in 8 % of the cases.

In Phase 2 of the study, 24 of the 34 CPS workers who had reported fictitious allegations of child sexual abuse were selected and placed in the “False Reports” subgroup while 24 of the 54 workers who had reported no false allegations were placed in the “True Reports” subgroup. The participants were asked to approximate what percentage of any 100 children making allegations of sexual abuse would lie about being abused. Results indicated that the workers in the False Reports subgroup expected more children to make false allegations (12.2 %) than those in the True Reports subgroup (5.2 %). The 24 CPS workers in the False Reports subgroup were also interviewed about details of the false reports of abuse in a sample of 29 cases and were asked to explain their judgment that the allegations of abuse were false. The participants judged the allegations to be false in a majority of the cases because the children had retracted their statements. They also determined that 14, or almost half of the children’s accounts, lacked credibility for reasons other than recantation. This consisted of unsubstantiated claims of sexual abuse made by the children in the past, statements that included fantastical and implausible details, insufficient amount of details for the child’s age and developmental level, inconsistencies in the statements, presence of contradicting evidence as well as absence of supportive medical evidence, lack of fear toward the perpetrator , and, in one case, the passing of a polygraph test by the alleged perpetrator. However, again, legitimate questions can be raised about the validity of some of these criteria, and it was unclear how many criteria were used in an individual case or how multiple criteria were combined.

Again, this study did not report the criteria for determining falsehood in each individual case, although they did attempt to uncover some of the criteria that were generally used. Like much of the research described above, this study relied on professional judgment that again may be prone to antecedent bias. Again, most of the criteria described, for example, complexity of statement, polygraph results, absence of medical evidence , etc., have not been shown to validly indicate that the claims are fictitious. Additionally, children have been known to falsely recant their statements, so retractions may also not be valid indicators of fictitious allegations . Finally, the authors provided no methodology for determining the child ’s intentionality. Thus, the study did not provide evidence that the child knowingly stated a falsehood, i.e., had lied.

The Public Health Agency of Canada conducted a large-scale incidence study of reported child abuse and neglect and published its major findings in 2003. The agency obtained data on 217, 319 child maltreatment investigations from 63 child welfare services areas across Canada (excluding Quebec) in five areas of interest: physical abuse, sexual abuse, neglect, emotional maltreatment, and exposure to domestic violence. Of these investigations, 47 % were substantiated (103,297). The findings excluded cases that were investigated only by the police, and reports that were screened out (that is, never investigated), either because of insufficient information about the child’s location or because they weren’t considered “to be within the defined mandate of the child welfare services” (p. 19). The investigation was the basis for judging whether reports were eventually substantiated, suspected, or unsubstantiated. Reports were considered “substantiated” if there was additional evidence corroborating the abuse, “suspected” if there was not enough evidence to substantiate maltreatment but in which maltreatment couldn’t be disregarded, and “unsubstantiated” if there was more evidence contradicting the abuse. Unsubstantiation was determined by the investigation worker and did not connote that the report was malicious. There were 12,468 child sexual abuse investigations conducted, of which only 21 % of these reports were considered substantiated (2935), while 15 % (1702) were suspected and 64 % (6244) were unsubstantiated. Caution is recommended when interpreting these findings, as children and other sources of referral were all included in the different categories. Of the 6244 unsubstantiated reports, most were considered to be non-malicious while a small number were deemed malicious (9 %) and the intent was unknown in 1046 (16.8 %). Once again there is no indication whether it was children or other sources of referral making the malicious reports. Three percent of all referrals (for all types of maltreatment) were unsubstantiated malicious reports by a child; unfortunately the study does not tease apart malicious reports by primary category of abuse.

The study presents a number of problems that limit interpretations of the findings. First, it does not differentiate between child sexual abuse allegations made by a parent, teacher, police, etc., and those made by the child. The numbers indicated that a relatively large number of reports were unsubstantiated; however, it is not known that children made those reports. Also unsubstantiated does not mean a false report, let alone a false report related to a child’s lie. Second, while the study does report rates of unsubstantiated malicious reporting by a child (this would more closely meet the criteria of a lie as “knowingly stating a falsehood”), the study lumps together all forms of abuse and does not indicate what percentage of unsubstantiated malicious reports were due to allegations of sexual abuse. Third, the findings also fail to report how “maliciousness” was defined. Fourth, the Public Health Agency of Canada relied on the professional judgment of the investigation workers to determine whether the abuse was substantiated, suspected, or unsubstantiated. This again introduces bias in such a judgment. Finally, there was no explication of the validity of the criteria used to determine that a report was substantiated, thus concerns about false positive rates are valid.

Studies Involving the Child’s Statement

Goodwin, Sahd, and Rada (1978) reviewed 46 cases of alleged child sexual abuse they had encountered in their work at a child abuse agency as well as an undisclosed number of cases from professionals working at other agencies in the Albuquerque area. All alleged abuse was perpetrated by a family member or someone living within the family. The authors found that of the 46 cases only 1 was a false accusation made by a child (2 %). This case involved a 13-year-old child who began exhibiting behavioral problems after her mother remarried. The girl had run away from her home and sought shelter from a friend whose father was a policeman. When questioned about why she ran away, the girl disclosed that she had been sexually abused by her stepfather. She later revealed that she had made up the story after reading about incest in a book. Two of the cases (4 %) were deemed false retractions of a true accusation made by a child. The two sisters ages 11 and 8 had run away and made claims of physical abuse. When those claims were investigated, one of the sisters also revealed that sexual abuse had taken place. A medical examination indicated that the older sister had “a ruptured hymen and a wide vaginal canal.” In a subsequent interview , the sisters recanted their story, calling it a hoax and revealing that they had been coached to make false allegations by an older girl. One of the sisters refused to provide more information about the hoax while the other cried and confessed that the mother had made up the retraction.

The authors only provided the child ’s statement as a means of verifying that the child knowingly lied about being sexually abused, e.g., in the case of the girl who admitted fabricating the story after reading about incest. Additionally, none of the methods utilized to determine that the child lied (either by stating that the sexual abuse took place when it didn’t or by recanting a true allegation) can be taken as conclusive evidence that lying occurred. In the case where a child made a false accusation, the criteria for concluding that the child’s claims were in fact false consisted of (a) the general circumstances of the initial outcry and (b) the child’s subsequent statement retracting the allegation. In the two cases where the study concluded that the children made false recantations , the criteria for establishing that the children’s recantations were truly false were limited to the children’s statements and some medical evidence that may or may not be indicative of sexual abuse. In addition, conclusions from this study are limited by sampling technique as there is no reason to believe that the original sample was representative. Finally, no methodology was used to establish that the child was knowingly stating false information.

Studies Involving Surveys of Professional Judgments and the Child’s Statement

Jones and McGraw (1987) examined the rates of false reports of child sexual abuse and features of fictitious reports in a two-part study. Part One reviewed reports of suspected child sexual abuse in 1983 to the Denver Department of Social Services (DDS) (N = 573). The Sexual Abuse Team of Denver DDS placed each report in one of five categories: reliable accounts, recantations , unsubstantiated suspicion, insufficient information, fictitious reports by adults, and fictitious reports by children. These five classifications were assessed by the researchers to ensure validity . Results indicated almost half of all reports (49 %) were reliable. Recantations, defined as reliable accounts that were taken back by the child under duress, made up 4 % of reports (N = 25). Insufficient information was provided in 24 % of reports (N = 37), while unsubstantiated suspicion made up 17 % of the cases (N = 96). Fictitious reports were made by adults 5 % of the time (N = 26) and only 1 % of accounts were deemed false reports made by a child (N = 8). The latter were judged to be fictitious if they contained deliberate falsification, misperceptions, or an adult coaching the child to make a false report.

Part Two of the study focused on establishing the validity of statements in 21 fictitious cases reported to the Denver DDS between 1983 and 1985, of which five were reports made by children, nine by adults, and the remainder were mixed cases in which it was not possible to determine who had made the initial allegation. Accounts indicated that four of the five children making false allegations had been sexually abused in the past and were currently suffering from PTSD symptoms while one was currently involved in a custody battle. The authors used the following criteria for determining the veracity of the children’s statements: presence of explicit as well as unusual and distinguishing details, age appropriateness of the child ’s’ words and sentence formation, perspective, emotion expressed, psychological response , pattern of abuse and elements of secrecy due to coercion or threats. Supporting features such as family history, child’s behavior, disclosure , statement to other people, consistency of the report, use of toys and other play materials, knowledge of sexual anatomy and function, and the presence of other children that may have been part of the abuse (e.g., as victims or witnesses) were also investigated as factors taken to provide further evidence for/against the truthfulness of the child’s statement. Additionally, the quality of the investigative interviews was considered and the authors determined that in 8 of the 21 cases no interviews were conducted, while of the remaining interviews only two met the adequacy criteria (length and developmental level of interview , exclusion of leading questions and of anatomically correct dolls). An evaluation of the children’s statements according to the criteria described above indicated that the false statements lacked emotion as well as distinguishing or unusual details, descriptions of threats, and the perspective of the child. The authors noted that the absence of emotion in the statements may at times be a symptom of unresolved PTSD. Contrary to what the authors anticipated about the amount and content of details in the fictitious reports (i.e., that fictitious reports were more likely to have insufficient details), a large number of details were present in the children’s false reports. The authors hypothesized that the information provided was a result of the child being coached by a parent or the child’s previous victimization. Finally, the researchers cautioned that the presence or absence of a particular feature does not make for a false report; rather, it takes multiple such features to be able to distinguish truthful reports from fictitious ones. However, they fail to specify the exact weighting of these or how many factors need to be present for such a determination. In addition, they failed to provide any information about the inter rater reliability of this key judgment.

Findings of the study are restricted by the lack of valid methods to reasonably decide that the children’s claims were in fact false. The authors did not specify a way to determine whether the child had knowingly made a fictitious claim, or if he or she truly believed the claim to be legitimate. Additionally, while the first part of the study took into consideration multiple types of fictitious accounts (recantations as well as false allegations declaring that the abuse did occur), it left out cases in which the child falsely denied that the sexual abuse took place. The second part of the study ignored all instances of lying which did not include a child claiming that the abuse occurred when in reality it did not. Furthermore, in both parts of this study “fictitious report” was defined as a report regarded by a professional to be false. Reliance on professional judgment may lead to a misestimate of true rates of fictitious accounts, even in cases where support was provided for the professionals’ decisions. Again, no assessment of the child’s intentionality was made.

Green (1986) assessed 11 cases of alleged child sexual abuse referred to the author (a psychiatrist) in the context of child custody evaluations. Results indicated that 4 of the 11 children (35 %) had falsely accused their fathers of sexually abusing them. The first case illustrated a little boy who disclosed to the author that he had seen his father ejaculate. The author concluded that child’s narrative lacked emotion and that his interactions with his father were positive, except for when the mother was present, when the child would behave in an angry and hostile way toward the father. The second case portrayed a mother who brought charges of sexual abuse against her ex-husband after her daughter came home with bloodstained underwear. The child reported that her father had rubbed against her, but later recanted her allegation and stated that she was only trying to please her mother and to stop her from the repeated questioning about alleged sexual abuse. The third case involved a 4-year-old boy whose mother suspected had been sexually abused after the child allegedly played a sexualized game that he claimed he learned from his father. The child later retracted his story, declaring that he had only made the allegations to stop his mother’s persistent inquiring. The fourth case depicted the maternal grandmother of a young girl who brought the child to be examined after the child protested to going to visit her father and complained of rectal and vaginal pain. The author decided that the alleged sexual abuse did not occur as evidenced by the child’s warm interaction with her father, lack of signs and symptoms of sexual molestation, and the pediatrician’s confirmation that the child had a chronic irritation on her bottom not due to sexual abuse. There was no mentioning of whether the child made any of the allegations of sexual abuse herself, or if they were all brought up by the grandmother.

The author relied on the child ’s statement to establish that the children in fact knew they were lying about the allegations of sexual abuse , for example, when they admitted to “making up” the stories to terminate their mothers’ questioning. The following criteria were used to determine that the sexual abuse allegations were fictitious: spontaneous disclosure without negative affect, child use of sexual terminology, discussion of the abuse by the child after checking-in with the mother, confrontation of the father by the child in the mother’s presence, positive interactions of child and father, paranoid and hysterical mothers who brainwashed their children into making the accusations, and signs and symptoms of child sexual abuse (e.g., PTSD symptoms).

Corwin, Berliner, Goodman, and Goodwin (1987) critiqued Green’s method of judging whether child sexual abuse allegations were true or false. Specifically, they criticized his use of a psychoanalytically derived technique that lacked empirical support, employment of his own clinical experience and anecdotal case reports and limited sample size. Additionally, the authors disagreed with Green’s judgment of the case of the 4-year-old boy presented above as later evidence revealed that the sexual abuse had in fact occurred. We agree with the authors that Green’s methodology was seriously flawed as most of his criteria for determining that the alleged abuse was false were not supported by the research. This study is a prime illustration of the dangers of relying on professional judgment, especially when basis of that judgment is Freud’s controversial theory regarding sexuality. In addition, it is clear that Green’s sample was not representative.

Benedek and Schetky (1985) presented 18 cases of alleged child sexual abuse they had encountered in psychiatric practice, ten of which they judged to be false (56 %). All cases were reviewed as part of child custody evaluations. Care must be taken when interpreting Benedek and Schetky’s findings due to the small and highly unrepresentative sample utilized by the authors. Furthermore, since it was not specified whether it was children or adults making the false allegations , it is unclear how many children, if any, made false allegations of sexual abuse . Therefore it cannot be determined how many children made false accusations of child sexual abuse. Due to this omission , it cannot be determined if there were any criteria utilized for establishing intentionality, namely whether the child knowingly and intentionally stated a lie.

The authors did, however, employ a variety of methods to assist their professional judgment of whether the child sexual abuse allegations were true or false. First they evaluated the child’s ability to distinguish fact from fantasy and assessed at the developmental appropriateness of the language used by the child. They cautioned that using precocious sexual vocabulary isn’t necessarily indicative of child sexual abuse , and that such language may be a result of the child having been “sexually overstimulated” by adults (i.e., witnessing parents interacting in a sexualized manner with each other) or coached by one of the parents. Next they evaluated potential “brainwashing” by the nonoffending parent. They also employed children’s play and drawings in the assessment of CSA under the assumption that this would help the children both in their disclosure of and coping with the sexual abuse. Additionally, they assessed for preoccupation with sex and displays of seductive behavior, which the authors once again indicate may be evidence of sexual abuse or of “sexual overstimulation” by adults. Finally, they engaged in direct questioning, evaluation of both parents, observation of parent–child interactions, and collateral information.

While the use of multiple methods of assessment is recommended for analyzing child sexual abuse allegations , many of the methods listed above have not been shown to accurately categorize whether child sexual abuse has occurred or not. The authors themselves admitted that precocious sexual language, preoccupation with sex, and seductive behavior are not always indicative of child sexual abuse. The use of any type of drawings and dolls are not supported by the literature, and sexually anatomically detailed dolls are particularly problematic as research examining their use (e.g., Elliott, O’Donohue, & Nickerson, 1993) indicated that a large number of nonabused children do engage in sexualized play with the dolls, increasing the risk of false positive identifications. Lastly, the accuracy of any observations and evaluations of the parents may be compromised by a variety of factors (e.g., parental stake in the outcome of a child sexual abuse investigation ). There is also no indication for what behaviors, histories, and disorders are good indicators of CSA and should be the focus of assessment (i.e., criminal record in alleged offending parent).

False Denials of Sexual Abuse and Recantations

None of the studies that we reviewed specifically looked at the rates of false denials of sexual abuse , that is, that children knowingly and falsely denied being sexually abused. Nevertheless, some of the studies do report rates of denial in cases substantiated by various agencies like CPS, and we will refer to those as false denials of sexual abuse. A large limitation of the research conducted on the topic is that none of the studies assessed for intentionality, and most utilized professional judgment as their criteria for evaluation of the child ’s statement validity . Because of this, we have only chosen to briefly review a few of the studies and found that false denials ranged from 2 (Malloy, Lyon, & Quas, 2007) to 72 % (Sorensen & Snow, 1991) and those of recantations from 4 (Bradley & Wood, 1996) to 27 % (Gonzalez, Waterman, Kelly, McCord, & Oliveri, 1993). Table 15.1 provides a more comprehensive list of published studies.

Table 15.1 Recantation and Denial Rates from Child Clinic Studies

Sorensen and Snow (1991) proposed a disclosure process of denial that developed in four stages: denial (initial statement indicating that no sexual abuse occurred), disclosure (tentative—acknowledgement of sexual abuse or active—personal admission by the child of being sexually abused), recant (retraction of the disclosure of sexual abuse), and reaffirm (reassertion that the abuse did in fact occur). The authors evaluated 630 cases of alleged child abuse they had encountered in their practice and selected 116 cases that they judged fit their proposed disclosure process. All of these cases were considered substantiated by one or more of the following: offender’s confession (80 %), conviction of offender (14 %), and substantial corroborative medical evidence (6 %). Results indicated that a majority of the children initially denied being sexually abused (72 %) and that most of these denials took place when the children were interviewed by a parent or other adult figure or in the context of a forensic interview . Of those children who initially denied being the abuse, 7 % went on to make active disclosures while 78 % provided tentative disclosures. Eventually 90 % of the children had gone on to make active disclosures. Children recanted their previous allegations in 22 % of the cases, and of those 92 % later reaffirmed the abuse.

Lawson and Chaffin (1992) evaluated false negative disclosures, defined as cases in which sexual abuse occurred but there was no verbal disclosure by the child , in a sample of 28 children aged 3 to puberty diagnosed with one or more STDs. The authors found that 12 children (43 %) provided a verbal disclosure during an investigated interview conducted by a social worker, while 16 (57 %) provided no verbal disclosure. The latter group fits the authors’ standard for false negative disclosures of sexual abuse due to the supporting medical findings of the STDs. Caregiver’s level of supportiveness was associated with disclosure by the child, given that 63 % of children with caregivers deemed supportive disclosed while only 17 % of children with caregivers considered unsupportive disclosed.

Bradley and Wood (1996) assessed 234 cases of CSA validated by Protective Services. About half of the cases (52 %) met Sorensen and Snow’s (1991) criteria for inclusion (medical evidence , conviction of offender and offender’s confession). Results indicated that of the entire sample, 13 of cases were denials (6 %) and 8 were recantations (4 %). Similar results were found when only the cases meeting Sorensen and Snow’s inclusion criteria were analyzed. Since the cases were considered validated, we can understand the percentage of denials as that of false denials , although we recognize that the accuracy of this number is limited by the accuracy with which Protective Services validated the cases. The authors determined that child ’s mother had played a large role in the child’s recantations in five of the eight cases through repeated pressuring of the child to take back the allegations of abuse .

Gonzalez et al. (1993) examined recantation rates in a sample of 63 children who had disclosed sexual and ritualistic abuse after attending preschool. The 63 children were evaluated by psychotherapists, and all disclosures were made in the therapy sessions. Results indicated that 76.2 % of children disclosed CSA in the first month of therapy, and of those 27 % recanted their allegations of sexual abuse, but most (88 %) later reaffirmed the statements. The therapists identified events related to system response (e.g., having to tell police, testifying in court) and events related to parent–child variables (e.g., parental pressure) as factors possibly associated with recantation. While the authors don’t label the recantations as “false” in this study, the overall sentiment is that the children falsely retracted their statements due to outside pressures, and that recantations should not be associated with false allegations (i.e., the child retracts his or her initial statement because that statement was fictitious) but rather seen as “a phase within the disclosure process for some children” (p. 288).

Malloy et al. (2007) analyzed rates of recantation of child sexual abuse allegations in 257 substantiated cases of CSA. Disclosures of sexual abuse were drawn from multiple formal (conducted by a professional, e.g., law enforcement personnel) and informal interviews (conducted by nonprofessionals, for example parents). Results indicated that in five cases (2 %) children never disclosed abuse. Because those cases were considered substantiated and there had been an attempt made during the interview to discuss sexual abuse, it can be concluded that this percentage to be demonstrative of false denials of CSA. Recantation occurred in 23.1 % of the interviews, at times within the same interview. Rates varied based on type of interview (formal vs. informal), informal interviews eliciting a slightly larger number of recantations than the formal ones.

As mentioned above, there are no criteria presented in any of the studies that indicates that any of the authors evaluated for the intentionality of the children’s statements concerning the sexual abuse . Thus, again, it cannot be said that the child lied about being sexually abused as knowingly stating a falsehood is required for this determination. Criteria for determining falsehood, that is that the children falsely denied or recanted sexual abuse even though the abuse did take place, varies from professional judgment (e.g., in the studies on disclosures and recantation s in therapy sessions) to a consideration of medical findings to the offender’s confession and/or conviction. Some of these measures are more valid than others. For example, if a medical examination finds that a 5-year-old child is infected with gonorrhea, one can safely infer that sexual abuse must’ve taken place for the child to contract a sexually transmitted disease. However, studies relying on the professional judgment of therapists, social workers, etc., introduce too much observer bias, therefore reducing the accuracy of the reported rates of denials and recantations . If one cannot accurately determine whether the abuse did in fact occur or not, one cannot accurately assess the recantation and/or false denial rates.

Discussion and Conclusions

Because of methodological limitations of existing studies, we reach the Socratic conclusion that we do not know the rates at which children lie about sexual abuse . That is, we know neither how often children lie about being abused when they have not been, nor do we know how often children lie about not being abused when they in fact have been. The principle methodological shortcoming that prevents such conclusions involves a lack of a valid method to determine whether or not the child is intentionally stating a falsehood. Thus, present studies are more relevant to the question of the frequency of false reporting as opposed to the question of lying —although again due to lack of representative samples and sound assessment of historical accuracy, even conclusions about this are problematic.

Studies, with all their methodological flaws, generally reported low rates of lying and false reports. There is some evidence to suggest that lying may be more associated with older children than younger children—which is interesting as younger children have been shown in the literature to be more suggestive (see Chap. 5 in this volume) and thus there may be differences based on age on pathways to false allegations . There is also some evidence to suggest that lying about abuse not occurring may be more common than lying that abuse did occur. However, again, given the significant methodological limitations of studies reviewed, these conclusions are very tentative.

However it is also notable that studies vary tremendously on their criteria used to determine historical accuracy. Some simply use professional “judgment” and usually even fail to explicate the details of this judgment. In addition, the studies reviewed fail to show the interrater reliability of these judgments. This metric would be useful as reliability sets a constraint on validity (Haynes, Smith, & Hunsley, 2011), that is, if any lack of reliability indicates a limitation on validity. Other studies use the confession or conviction of the offender; however, we know that people have falsely confessed of crimes, and that a conviction does not guarantee that the abuse took place as people have been falsely convicted of child sexual abuse . Other studies use a variety of criteria, many of which have not been shown empirically to be valid indicators and most studies also fail to show how multiple criteria were combined to make ultimate judgments. In addition, no study examined the subcomponents of the child’s claims individually to determine which were false and which were true. All studies took a rather global perspective and either judged all the children’s claims as true or all as false.

This knowledge gap is important because it calls into question certain lines of reasoning that may be used in actual cases. This knowledge gap certainly questions the reasoning of advocacy organizations such as Believe the Children and their claims that children never lie. This strong claim clearly has not been established in the empirical literature. However, it also calls into question more nuanced claims that use the reasoning that because the rate of children’s lying is trivially low (say 1 or 4 %) and that therefore these very low base rates suggest that some particular case of child abuse ought to be believed. Conversely, our review of the literature suggests that the same sort of argumentation is flawed regarding denials of abuse, i.e., that since we do not know the rates of lying about this, we also have to be cautious of the use of percentages in our arguments regarding this. This is particularly true in actual court cases as there have been no studies of the rattles of lying in actively adjudicated samples. However, we must also quickly say that there is no evidence to suggest that lying is a highly frequent phenomena—certainly there is no evidence that the majority of even a sizable minority of allegations are lies .

This review suggests that more research is needed. Future research ought to seek representative samples as well as to examine samples involved in judicial proceedings as there are reasons to believe that the rates of lying in these samples may differ from those in the general population For example, one reason why these rates may be different is that there may be an increased motivation for falsely accused individuals to adjudicate rather than accept a plea bargain. In addition, future research should examine special samples of children, for example, an interesting partition may be children with a history of lying about other issues or children with significant psychopathology; or children who have been threatened by their possible perpetrator . In addition, research ought to more carefully handle the question of the child ’s intentionality as well as more carefully address the question of historical accuracy. Finally, we suggest a more molecular approach be used and accuracy judgments be made about individual components of the child’s allegations , e.g., who was the perpetrator, how many times this occurred, etc., as these are key components of the child’s allegations and have important consequences, e.g., the jurisdiction and the number of counts.

The implications for forensic interview ing are also unclear. The research suggests that professionals do not have a better rate of detecting lying than nonprofessionals. In addition, there are not valid gross indicators of lying (gaze, blushing, etc), especially as these may also occur when discussing sensitive matters like sexual abuse . Thus, one caution would be to take a skeptical stance of interviewers who come to strong conclusions about lying in their interviews . Finally, it is not clear what methods ought to be adopted in the forensic interview . Most protocols incorporate whether the child knows the difference between a truth and a lie, but knowing this differences does not mean that the child will not then tell a lie. Some protocols attempt to emphasize the importance of truth during the interview but again, there is no evidence that this has any effect on increasing the probability of truth. Intentionality is a notoriously difficult construct for an outside observer to accurately assess and thus will always present a conundrum in forensic interviews .