THE EXPERT WITNESS: Exploring alternative hypotheses and obtaining reliable forensic evidence in CSC cases - The most common errors in forensic interviewing


By Michael G. Brock

"Scientific Method: principles and procedures for the systematic pursuit of knowledge involving the recognition and formulation of a problem, the collection of data through observation and experiment, and the formulation and testing of hypotheses"

-Merriam-Webster Dictionary

"As one brief filed inDaubertsuggested, 'Scientific methodology today is based on generating hypotheses and testing them to see if they can be falsified; indeed, this methodology is what distinguishes science from other fields of human inquiry.i,ii'"

-Michael D. Greeniii

"I cannot give any scientist of any age better advice than this: the intensity of the conviction that a hypothesis is true has no bearing on whether it is true or not."

-Peter Medawar, "Advice to a Young Scientist"

Alternative hypotheses are mentioned 23 times in the Michigan Forensic Interviewing Protocoliv (Protocol), which is law for forensic interviewers who are government employees in this state.v The research clearly shows that an interview that includes only one possible finding will inevitably reach the conclusion that it is Consequently, the first principle of forensic interviewing is that the forensic interview is hypothesis testing rather than hypothesis confirming.vii

I have reviewed numerous forensic interviews done for courts, most of which were performed at a center set up by the county for this purpose, such as Care House in Oakland and Macomb Counties, and Kids Talk in Wayne County. These are facilities established for the purpose of providing a scientific foundation for the introduction of a child's testimony into evidence. Their guide for conducting forensic interviews of children is the Protocol, but I don't recall ever having seen one that seriously tested even one alternative hypothesis, with the possible exception of an occasional tag question at the end of the interview: "Did anyone tell you what to say here today?"

The lack of alternative hypotheses cause the interviewer to unconsciously and subtly or not so subtly lead the child to the expected answer-that she has been abused. If children feel that the adult asking them questions expects a certain answer, perhaps because a question is repeated more than once, the interviewer is an authority figure, or the interviewer has introduced a topic that the child has not yet raised, the child is likely to tell the interviewer what they expect to hear.

The Protocol (like any protocol based on the research) and underlying forensic literature also state that the best evidence is that offered through a free narrative.viii A child who has been abused can provide a reasonably cohesive narrative with some detail. Yet the Protocol also states that the most common error in a forensic interview is not asking for, or allowing a child to present a free narrative.ix This is often because the interviewer does not allow sufficient silence for the child to take the lead, as it were, in the interview process. It suggests that silence, "can exert a subtle but gentle pressure on the child to respond."x (P. 15) That is to say, a subtle pressure on the child to lead the interview, a role that is somewhat uncomfortable for both the interviewer and the interviewed.xi

Moreover, by following up too early with clarifying questions, the interviewer helps children build their stories with short answer, yes/no, or either/or questions, or sometimes literally tells the story themselves with leading questions and their own additions to the child's responses. A child offered an either/or question often feels obligated to answer the question by choosing one of the alternatives, whether or not the alternative is actually true. Yes or no questions can be tricky also, because the child will often feel pressured to provide the answer they believe is expected by the interviewer. Leading questions present information the child has not yet demonstrated knowledge of and may not possess, or suggest "correct" answers.

Not included in the Protocol, but mentioned to me in an email by its author, Dr. Debra Poole, is that with teenagers, the problem is not as much one of the child being led, as the possibility that they are lying. In my experience, liars typically contradict themselves. This is, I think, a universally accepted truth (Not that everything universally accepted as truth is necessarily true, but this one happens to be.). The best indicator that anyone is lying is the presence of contradictions in the story they tell about the incident. The contradictions may be within the context of a single interview, or from one report of the incident to the next. But just as with small children, when forensic interviewers ask short answer and leading questions rather than having interviewee construct a story through their own narrative, the interviewer is actually helping to tell the story, which is the antithesis of what the interview is supposed to accomplish.

In the following scenario of a case which recently resulted in a not guilty verdict from the jury,xii the interviewer could have considered these alternative hypotheses, keeping in mind that suggestive interviewing methods can produce false positives.

The background is that two sisters (Megan 15, and Claire 13) filed identical complaints of a single incident of rape against a former boyfriend of mom's. The incidents allegedly occurred approximately five years earlier, when the girls were eight and ten. At the time of the allegations, the older child was pregnant and had been missing school due to morning sickness. The younger child has gone to see the school social worker at the middle school with the stated goal of having her middle school call the high school to explain and excuse her sister's absence, and to keep her mother and sister out of trouble. This concern was clearly stated in the forensic interview with the younger child, but never followed up on.

An alternative hypothesis would have been that the child might be lying in an attempt to portray her sister as a victim, rather than as truant and sexually active.xiii,xiv A question offered by the Protocol as a way to explore this possibility is, "How did you feel about (mom's boyfriend)?" It turns out they didn't like him, which they admitted. This fact provides an additional motive, but there is probably no specific question that will reveal a person to be lying. If they are motivated to lie, they are not going to admit it. Still, it is something the interviewer has to take into consideration when weighing the child's allegation. If the child's story is not very consistent or probable, the interviewer needs to relay that information to the criminal investigator and a decision made regarding whether or not to move forward. Once the decision is made to charge the suspect, there is little possibility that the case will not go to trial.

The conversation Claire had with the social worker is not known and has been kept confidential as privileged communication between a therapist and client, but the allegations of abuse emerged from the interview. Another alternative hypothesis would be that that the allegation might have been the result of suggestive interviewing and confirmatory bias by the initial interviewer (the school social worker) who reported, not that the children had alleged rape, but that they had been raped, their mother had not protected them, and they were being inadequately fed. The Forensic Interviewer could have explored the circumstances under which the children were interviewed by the school social worker with a question like, "Can you tell me about your discussion with the social worker? How did you get to talking about abuse?" The goal would be to find out if the rape allegations were volunteered or were elicited by suggestive questioning. This alternative hypothesis was never explored.

One thing that appears frequently in forensic interviewers are statements that the interviewers are there to keep the child safe. This in itself displays an interviewer bias. They are supposed to be scientists who are committed to helping the fact-finder discover the truth, even when that truth is that the client is lying or being coached. If they begin with the belief that the child is a victim, they are already possessed of a confirmatory bias,xv which will influence both the methods they employ and the results of their investigation.

Alternative hypothesis testing of the scientific method conflicts with the politically correct view of believing the victim. How do we know she's a victim? She said so. The circumstances don't matter. An alternative hypothesis of the older child's allegation in this case would have been she might be feeling the need to go along with her sister's allegation in order to deflect blame, to get out of trouble, or excuse behavior that she felt made others look down on her. At one point in the interview, the older child expresses this concern. Was her fear of reputation such that it would cause her to make such an allegation? Keep in mind: teenagers do not understand the consequences of these actions for those they accuse. They are mentally children in adult bodies.

The older child said at one point during the forensic interview that when she was raped at the age of ten, she thought others would not look at her the same way afterward. These are not the thoughts of a ten year old during the process of being raped. They are the thoughts of an older child who is concerned about her reputation. Did the girls talk to each other between the times that Megan and Claire made their allegations? Did Claire's allegation put the idea in Megan's head or vice versa? Why did they both wait until years later to make this disclosure? That one child may have been suggestible to the influence of the other is an alternative hypothesis.

Protocol says to question whether the incidents could have happen in the way they are described. Could the assaults have happened the way the girls said they did? Megan initially said that her mother and sister were sleeping in the living room, then changed to saying that her mom was sleeping in the other bedroom. Despite contradictory statements that her attacker held his hand over her mouth, she claimed that she was screaming. Would it have been possible for her to be yelling and not have been heard by her mother? Is this reliable evidence?

Megan, when asked a leading question, "Did he say anything to you, [while he was raping you]? Replied that he told her not to tell anyone or he would do it again (a story which she later changed to, "He said if I told anyone he would come back [in an ominous tone])." She then said she responded that she would wait until two or three days later to tell her mom, when Oliver would be gone and couldn't dispute it. Are these the thoughts, actions and conversation of a child who is being raped? Is this reliable evidence?

The children said they didn't like Olivier or anyone dating their mother, and that they knew he was cheating on their mother. Neither of the girls mentioned pain or psychological trauma when relating the rape incidents during the forensic interview. Megan denied any injury when questioned by the judge in the preliminary hearing, and Claire never mentioned pain until the idea was put in her head by the prosecutor at the preliminary hearing. Is it possible that these pre-pubescent girls would not have both experienced physical and psychological trauma when being violated by an adult male penis? Is this reliable evidence?

Claire said that she at first thought what she experienced was a dream, then discovered it wasn't years later when having a sex talk with her sister Megan, a talk that Megan denied ever took place. Is a dream that becomes reality only through a non-existent conversation reliable evidence? Is it even possible that neither child would have experienced pain and soreness at the time of the rape and at least for the following day? Is this reliable evidence?

Claire said that she told her mother she wet the bed. Is it possible that no one noticed the blood from the violation? That she did not notice the consistency of the semen? Is this reliable evidence? Megan had made a previous allegation of being molested or raped by someone else, which she told her mother about and later admitted was false. What was the reason for that allegation? Was it after her first sexual experience because she was feeling guilty? Feared she would become pregnant? An alternative hypothesis would be that the child could have made another false allegation for the same reason.

Is it reasonable that Megan would have told her mother about the one violation that didn't happen, and not told her about one that actually happened? Is this reliable evidence? First Megan said she didn't tell her mother, then she said she said she told her mother, but her mother didn't believe her. Mom said neither child made the allegation to her. How likely is it that they would have both told their mother they had been raped and that she would not remember either one? Questioned them more? Reported it to police? Discussed it with Olivier? Is this reliable evidence?

Much of what the children were saying contradicted itself, and/or contradicted a credible scenario of how an assault could take place and what the consequences, both physical and psychological would be. There is no foolproof way to tell if someone is lying. That is why polygraphs are not admissible. Consistency is the best way of determining if someone is telling the truth. This testimony is inconsistent. Is this reliable evidence?

At one point in the forensic interview with Megan, she asked the interviewer if she, Megan, had killed or raped someone, would she be brought in for a forensic interview, have to appear in court, or possibly be looking at jail time? The forensic interviewer could have explored as an alternative hypothesis whether the child might be feeling guilty about making a false allegation with a question like, "Why is that a concern for you?" "Can you tell me why you raised that question?" She could also have helped the child to understand that anyone found guilty of such a crime would not have a life worth living if he ever got out of jail. She could have stressed the importance of telling the truth in a legal proceeding. Instead, she tells Megan that it is doubtful she will ever be in that situation, and that the interviewer is not there to investigate her, she is there to keep the child safe.

That is not entirely true; the forensic interviewer is there as a scientific investigator whose purpose is to help the court reach a just decision. The notion or statement that the forensic interviewer is child advocate demonstrates interviewer bias in two ways: first, it suggests that she is part of the prosecution team rather than an objective investigator; and, secondly, that the possibility of the child or someone manipulating the child could have ill intent are not possible options. Two alternative hypotheses that should be considered in any forensic interview are that the child could be lying, and/or that someone could have programmed the child to make a false report.

It is a terrible crime that someone could injure a child in any way, but it is equally terrible to make a false allegation of abuse in order to get revenge, or to gain the advantage in a child custody dispute. Any family law attorney knows that this happens all the time. It is also common for children (especially teens) to make a false report of abuse to take revenge or get out of trouble. The fact that there are upsides, but seldom negative consequences to such behaviors makes them more attractive optionsxvi. A forensic interviewer who does not consider these possibilities, but who sees herself as part of the police/prosecution team intent on "cracking the case" will inevitably make serious mistakes.

This child's question suggests that she has no understanding of the consequences of her allegation on the life of the person she has accused. An alternative hypothesis could be that the child found it easier to lie about a serious matter because she really did not understand the havoc that such a lie would wreak in the life of an innocent person, and ultimately on society as a whole. The interviewer could have explored that hypothesis, with questions like, "Do you understand that this is among the most serious crimes a person can be charged with?, and that anyone convicted of such a crime is going to be in prison for a very long time, then be on the sex offenders list for the rest of his life?"

Both girls said the rapes happened only once. Sexual predators are repeat offenders, that's why we have a sex offenders' registry. Even with treatment they are rarely rehabilitated. Moreover, their mother didn't believe that Oliver had raped them, so what was to keep him from repeating the behavior as the older child said he had threatened to do? How reliable is this evidence?

At the preliminary hearing the judge noted that the children's testimony was inconsistent, but that did not dissuade him from binding the defendant over for trial because he was not convinced that it was improbable a reasonable jury could find the defendant guilty. He also decided, based on an unpublished case, that the jury did not need to hear testimony from an expert on the Protocol and the volumes of forensic research and publications that are its underpinnings (the life's work of its author Debra Poole and many others) because the Protocol was understandable by the average person. My experience has been that neither legal professionals, not those entrusted with applying this forensic tool sufficiently understand it, let alone a jury of non-scientists.

A defense expert is significantly limited in sexual assault cases because they cannot offer an opinion as to whether the event actually happened, or whether the child is being truthful. But taking the defense expert out of the equation entirely, means that only the prosecution expert-the forensic interviewer-is allowed to testify. Though the defense has a right to cross examine, they were not allowed to offer a scientific expert to controvert the testimony of the prosecution witness. Juries expect to hear testimony from forensic experts, and they put a lot of faith in such testimony. Legal professionals refer to this as the "CSI effectxvii."

Ultimately, defense counsel was able to present these inconsistencies in the children's statements from the forensic interview, through preliminary exam, to testimony at trial. However, I have seen cases where the only evidence permitted was the child's testimony at trial, and no discussion of the child's previous statements was allowed. Imagine a murder case with no physical evidence where the prior contradictory statements, or perhaps confessions by the defendant, were not allowed into evidence-just what the defendant said at trial. Would the prosecutor in such a case believe he had a fair chance to achieve a just outcome?


While writing this article, I was looking for the first edition of the Michigan Forensic Interviewing Protocol. My recollection was that the wording of the introduction was different, and that it included statements that among the goals of the Protocol was to present children's evidence in a "legally defensible manner" and also "to protect the rights of the accused," which are not in the current edition. I no longer have my original version, which was given to me by Beth Clark at a seminar on forensic interviewing, and was wrapped in a yellow binder, soI wrote to its author, Debra Poole, to ask if she might possibly have an email version of that protocol she could send me.

Dr. Poole wrote back, sending a version of the original Protocol. It contains the following quote: "The purpose of this protocol and training is to prepare local investigators to conduct competent child interviews which will reduce trauma to children, make the information gained more credible in the court process, and protect the rights of the accused."(emphasis added) This is on page five of the PDF, the "Introduction to the Introduction," as it were. I find it distressing, but not accidental that this language has been take out of subsequent versions. In the People v. William Watters case I testified that the evidence had been destroyed by the deliberate misuse of the Protocol. I wrote about it in an article entitled, Presumed Innocent until Proven Broke, Part II. Briefly:

"Initially, the police officer tried to follow the Protocol and to see if the child would provide a voluntary narrative that would confirm the allegation. When he was unable to obtain a narrative using proper forensic techniques, the police officer began to ask closed-ended, yes or no questions of the child:

'Did you see anything?'


'Did you feel anything?'


'Did you hear anything?'


'Did he say anything to you, ever?'


"At this point the police officer made a significant departure from Forensic Interviewing Protocol by asking the child a leading question which presumed that a disclosure of abuse had already been made. The next question was:

'How many times did this Happen?'"

After seeing the tape of this forensic interview, and having this breach of Protocol explained to them, my jury hung. In the retrial, Dr. Katherine Okla was retained as the expert and her testimony was limited because the judge ruled that the Protocol was self-explanatory and there was no need to have an expert explain its application to the jury. The appeals court upheld the verdict:xviii Two of the appeals court judges (Krause and Kelly) stated that, "The trial court concluded that the protocol itself was not difficult to understand, and allowing Dr. Okla to comment on what she believed to be specific violations thereof tread dangerously close to impermissible commentary on the credibility of the witnesses. We do not find the trial court's decision to be unprincipled."

The third judge (Shapiro) wrote a concurring opinion in which she stated that testimony was permissible because, "Such testimony goes to the reliability of the investigative process and not to the credibility of the child complainant. For this reason, the trial court's limitation of the expert's testimony to the mere existence of the protocols was overly broad and an abuse of discretion." So far, so good. But she goes on to say that the error was harmless because, "First, defendant's offer of proof suggested only a single deviation from the protocols. Moreover, as noted by the majority, the jury was not left to determine the truth of the allegations based on the taped interview alone. Both the complainant and her brother offered live testimony, during which they were subject to cross examination." The Michigan Supreme Court refused to hear the case.

What this decision essentially means is that the evidence regarding how the disclosure of abuse was obtained-which purports to be a scientific process-is irrelevant when the allegation is repeated at trial. Therefore, the fundamental principle of forensic interviewing-that the forensic interview is hypothesis testing rather than hypothesis confirming-is essentially irrelevant to what is produced by the ostensibly scientific process. In short, biased science practiced by those with a vested interest in the outcome of the process and devoid of the most basic principle of all science, hypothesis testing, is acceptable in a court of law. Testimony by a defense expert pointing out the biased application of ostensibly scientific principles is not. This is despite the research that shows once a child has been convinced something that did not in fact happen, has happened they cannot be then dissuaded, and will tell a convincing story of what they now believe to be a true event.xix

It goes without saying that this trend clearly favors the prosecution in cases where there a tendency of juries to presume guilt, and where the mere fact of a child stating she was abused is enough for most jurors to vote for conviction. Add to this the difficulty of obtaining timely discovery in the form of video or even (less reliable) transcripts, and the burden of proof decidedly shifts to the defense. Sadly, we have come full circle to the days of the 1980s' Daycare Hysteria in which dozens of lives were ruined based on honestly held beliefs held by innocent children as the result of suggestive and coercive interviewing techniques by well-meaning mental health professionals who's confirmatory bias was reflected in their relentless pursuit of a single hypothesis-that the children had been abused.

God save us from the well intentioned.


i National Academies Press: OpenBook National Academies Press: The Age of Expert Testimony: Science in the Courtroom, Report of a Workshop (2002), Chapter: 4, The Scientist's Role in the Courtroom,

ii Green, Expert Witnesses and Sufficiency of Evidence in Toxic Substances Litigation: The Legacy of Agent Orange and Bendectin Litigation, 86 Nw. U. L. Rev. 643 (1992), cited in Daubert amici curiae brief of Nicolaas Bloembergen et al.

iii Michael D. Green, Law Professor, Wake Forest University, School of Law, is a nationally and internationally recognized torts teacher and scholar.

iv State of Michigan Governor's Task Force on Child Abuse and Neglect and Department of Human Services Forensic Interviewing Protocol, Third Edition

v CHILD PROTECTION LAW Act 238 of 1975, 722.628 (4) (6)

vi Suggestibility, reliability, and the legal process, Robert Rosenthal, J.D., Development Review 22 (2002) 334-369

vii Page 1, Michigan Forensic Interviewing Protocol, "First, forensic interviews are hypothesis-testing rather than hypothesis-confirming (Ceci & Bruck, 1995). Interviewers prepare by generating a set of alternative hypotheses about the source and meaning of the allegations."

viii U.S. Department of Justice Office, Juvenile Justice Bulletin, Sept 2015, Child Forensic Interviewing: Best Practices: "Maximizing the amount of information obtained through children's free recall memory is universally accepted among forensic interview models as a best practice. Forensic interviewers should use open-ended and cued questions skillfully and appropriately to support children's ability and willingness to describe remembered experiences in their own words (Lamb, Orbach, Hershkowitz, Esplin, and Horowitz, 2007; Myers, 2005; Saywitz and Camparo, 2009; Saywitz, Lyon, and Goodman, 2011)."

ix Protocol, P. 16: "The most common interviewer errors are omitting the free narrative phase or shifting prematurely to specific questions."

x P. 15, Protocol

xi P. 16, Protocol

xii The names were changed to protect the parties involved, including the not guilty defendant.

xiii Alternative Hypotheses, P. 25, Protocol, "The child reported sexual abuse to cover for evidence of sexual activity."

xiv That the older child was sexually active was known to defense counsel, but could not have been brought up at trial because of the rape shield law. However, this is another example of how laws to protect the accuser keep the defendant from getting a fair trial.

xv Confirmation bias: In psychology and cognitive science, confirmation bias (or confirmatory bias) is a tendency to search for or interpret information in a way that confirms one's preconceptions, leading to statistical errors.

xvi Simply Psychology: B. F. Skinner - Operant Conditioning, by Saul McLeod published 2007, updated 2015 "Skinner is regarded as the father of Operant Conditioning, but his work was based on Thorndike's (1905) law of effect. Skinner introduced a new term into the Law of Effect - Reinforcement. Behavior which is reinforced tends to be repeated (i.e. strengthened); behavior which is not reinforced tends to die out-or be extinguished (i.e. weakened)."

xvii The 'CSI Effect': Does It Really Exist? by Honorable Donald E. Shelton "The bottom line is this: Our criminal justice system must find ways to adapt to the increased expectations of those whom we ask to cast votes of 'guilty' or "not guilty.'" NIJ Journal No. 259, March 2008, NCJ 221501

xviii opinions/appeals/2015/081815/60649.pdf

xix Truth and Consequences, Dr. Stephen Ceci,


Michael G. Brock, MA, LLP, LMSW, is a forensic mental health professional in private practice at Counseling and Evaluation Services in Wyandotte, Michigan. He has worked in the mental health field since 1974, and has been in full-time private practice since 1985. The majority of his practice in recent years relates to driver license restoration and substance abuse evaluation. He may be contacted at Michael G. Brock, Counseling and Evaluation Services, 2514 Biddle, Wyandotte, 48192; 313-802-0863, fax/phone 734-692-1082; e-mail, michaelgbrock@; website,

Published: Wed, Apr 19, 2017