Hsu Pei-ming Case

A disclosure process that can never be reconstructed

"Daddy, you wouldn't give your birdie to the puppy to eat, would you?"

According to the judgment convicting Hsu Pei-ming, on September 19, 2008, an eight-year-old girl with moderate intellectual disability (referred to in the judgment as "Girl A") blurted out the sentence above during a family outing, after seeing a puppy run under the driver's seat where her father was sitting. Startled, Girl A's mother pressed her further, and Girl A said that a teacher at school had also given her his "birdie" to eat. The mother then took Girl A to school to see her homeroom teacher, who showed Girl A a group photo of all the teachers in the school yearbook. Girl A identified teacher Hsu Pei-ming. The homeroom teacher promptly reported the matter to the sexual assault prevention center and the police, and the authorities focused their investigation on Hsu.

What exactly happened during the disclosure of this suspected sexual assault can never be reconstructed, because there was no audio or video recording. What Girl A actually said and how she said it, how her mother asked, how the teacher asked — none of this can now be confirmed or examined. The formal investigation stage that followed, however, was recorded.

A severely leading judicial interview process

The statements forming Girl A's identification of Hsu Pei-ming as the perpetrator came chiefly from the initial questioning at the police Women's and Children's Protection Corps (October 2, 2008) and the school gender equity committee's investigation (first interview on October 30, 2008; second interview on November 17, 2008). Statements from these inquiries were pieced together fragment by fragment into: "While administering the Wechsler intelligence test to Girl A in the resource classroom, teacher Hsu Pei-ming tied a towel over her eyes and put his genitals into her mouth." But the video records show that the questioning of Girl A was riddled with severe leading.

"Leading questioning" means that "the questioner inserts into the question details the interviewee never mentioned, or directly hints at the desired answer within the question." Leading can be seen throughout the police questioning and the gender equity investigation. For example, at the start of the police interview, Girl A kept playing with the anatomical doll and would not respond to the policewoman's questions. The policewoman eventually lost patience and, as Girl A undressed the doll and played with its genitals, asked: "Why do you keep playing with that!", "What is that?", "What is that organ called?" Girl A answered: "Eyebrow." The social worker present could not accept that answer and, pointing at the doll's genitals, asked: "Then what about this?", "What is this?", "What is this long thing?" Girl A answered: "Birdie." The policewoman pressed on: "Whose birdie have you seen?", "Have you seen anyone's birdie?" Girl A answered: "Teacher Hsu Pei-ming's."

Whenever Girl A's answers about the details of the assault did not match the adults' imagination, the adults would repeat the question, ignore her answer, or come up with what they considered a more plausible scenario and ask about that instead. For example:

At the police interview, to the question "Did teacher Hsu Pei-ming put his birdie into your mouth?", Girl A sometimes said "yes" and sometimes "no." The policewoman asked her: "Was your mouth open?" Girl A said: "No." The policewoman then retorted: "If not, how did he put it in?"

As for the scenario of Hsu covering Girl A's eyes with a towel, it was the social worker who first asked Girl A, "Did he cover your eyes?" — only then did Girl A say such a thing had happened.

When the policewoman asked about the towel's size, she thought the length Girl A gestured was too short to tie around the eyes and said, "How could something so small be tied?" — but rather than pursue the question, she and the social worker together used the doll for role-play, handing Girl A a towel and asking her to try tying it.

On top of the leading questioning, during the gender equity investigation the committee members also used "playing a game" to have Girl A describe the events. This made the contamination of the testimony even more serious, because it becomes impossible to distinguish whether Girl A thought she was playing a game or recounting something that really happened:

Committee member A: "Let's play a game now. Pretend you are Teacher Hsu, and the doll is little Girl A. Act out for me what Teacher Hsu did to little Girl A, okay?"
Girl A: "Okay"
……
As Girl A played with the doll's mouth, committee member A asked: "This is the mouth, right?"
Girl A: "This is my hand."

Here the two were clearly focused on different things, but committee member A, intent on asking about oral sex, continued: "Why is your hand digging at the mouth?" Since Girl A simply would not answer in the direction committee member A wanted, the member finally just said outright: "You told auntie just now that Teacher Hsu put his birdie into little Girl A's mouth, right? Did he?" But Girl A had never mentioned this — a plain instance of leading — and still Girl A did not react. Another member, B, later added: "Kept touching your mouth," and only then did Girl A nod.

From the question-and-answer at the police interview, one can see that the social worker and policewoman present had already fixed on the premise that Hsu Pei-ming had committed sexual assault, entertaining no other possibility. The social worker, having probably questioned Girl A beforehand, had a preconceived assault scenario in mind and kept steering Girl A toward telling it. The policewoman followed the social worker's leading, adjusting details of the scenario she found implausible. And in the gender equity investigation, the committee members used leading combined with game-play to guide Girl A into "performing" the assault once more. Through this process, Girl A gradually told the so-called "truth." Little did anyone realize that this "truth" was already contaminated — whether it was the truth of what happened, or the truth the adults believed in their hearts, can no longer be known….

A psychiatric evaluation that overstepped its expertise

After the police interview and gender equity investigation, the prosecutor — to determine whether Girl A suffered post-traumatic stress disorder (PTSD) from the assault — arranged for her to be evaluated at Kaohsiung Chang Gung Memorial Hospital on March 31, 2009. The evaluating physician's conclusion was: "no evident post-traumatic reactions." Despite this conclusion, the physician went on to interpret the result as follows: "Although Girl A experienced a sexual assault, given her limited cognitive capacity, and since the perpetrator used no excessively violent or coercive behavior during the assault, Girl A could not understand the violating nature and meaning of the event and therefore displayed no post-traumatic reactions. The perpetrator exploited the victim's disability-related vulnerability, lack of wariness and capacity for self-defense, and his own position to violate the child — his intent is deplorable and his conduct unforgivable."

The physician's task was to evaluate whether Girl A had PTSD, and his conclusion was "no" — that is within his professional judgment. But whether a sexual assault actually occurred, and what the perpetrator's state of mind and actions were, are judgments beyond the scope of a PTSD evaluation, and cannot be known from a single day's examination, testing, and interview of Girl A. In making those pronouncements, the physician violated the professional ethics of forensic psychiatric evaluation.

A guilty verdict built on flawed evidence

On the strength of an unverifiable disclosure process, severely leading interviews, and a psychiatric evaluation that overstepped its expertise, the court convicted teacher Hsu Pei-ming of an adult intentionally committing sexual penetration against a child by taking advantage of the child's condition, sentencing him to 5 years and 10 months of imprisonment….

In truth, the court was not unaware that the evidence in this case was full of flaws — defense counsel had raised them long before. But the hurdle the court could not get past was this: "If the defendant really did nothing, why would the little girl say he did?" The court thus held that a child with moderate intellectual disability would not lie, nor confuse reality with imagination. "Leading" was thus treated as a "necessary assistive measure"; contradictory or implausible statements were regarded as mere differences in detail that did not shake the basic fact that the girl said she had been assaulted.

An opportunity for exoneration: the latest research in forensic psychology

Questions of whether to credit witness testimony used to be treated purely as matters for the court's fact-finding. But the latest research in forensic psychology offers an opportunity to re-examine how testimony is obtained from children with intellectual disabilities. According to this research, the way a forensic interview of a vulnerable victim is conducted strongly affects the credibility of the testimony. Psychology can therefore assess, by evaluating how the forensic interview was conducted, how severely the victim's testimony was contaminated or improperly influenced.

Forensic psychology research shows that human memory decays over time, and new information becomes integrated into old — so-called "false memory" is in fact very common. False memories can arise when a person is asked to recall something and the request carries suggestive content or misleading information. They are especially likely when the person cannot really recall the event, or is encouraged not to question whether their memory is real. The renowned psychologist Elizabeth Loftus, drawing on her experience with wrongful conviction cases in the United States, has noted that in the 1990s many American patients, after receiving help from psychotherapists, began to develop strange and even unreal memories, giving rise to numerous wrongful sexual assault convictions. Loftus attributes this to therapists who assumed their patients had repressed subconscious memories and tried to "awaken" deep memories through imagination exercises, dream interpretation, and similar techniques — which instead produced false memories.

It is precisely from this understanding of memory's fragility that the United Kingdom and United States today place great weight on how forensic interviews of vulnerable victims are conducted — both to ensure the victim's testimony can be admitted by the courts, and to prevent wrongful convictions. A quality forensic interview capable of producing credible testimony follows clear procedural guidelines, with corresponding environmental conditions. Procedurally, the basic structure is: building rapport before substantive questioning; transitioning to substantive questions to learn the event as it originally appeared; investigating the event; and closing. Before substantive questioning, the child must be helped to understand the purpose and method of the interview, and told that they may say "I don't know" or "I don't remember," and may correct or object to the interviewer's statements or questions at any time. At the transition to substantive questions, the child must recount the events on their own, free of suggestion from the questioner. During the investigation of the event, open-ended questions should be used as much as possible, built on information the child has provided, to avoid leading or suggestion. The interviewer's basic stance should be to seek evidence for multiple rival, competing hypotheses — for example, that an assault did or did not occur, that the perpetrator might be A or might be someone else — rather than locking onto a single hypothesis. To avoid contamination, it is recommended that the child be interviewed alone; if another adult must be present, they should be kept from speaking. If an anatomical doll is to be used, it should be introduced only after the questions have been asked, and leading questions should still be avoided.

The police questioning and gender equity investigation in the Hsu Pei-ming case described above plainly failed these requirements for forensic interviews: the entire questioning was locked onto Hsu having committed sexual assault, with no competing hypothesis; a social worker present kept making leading suggestions; the transition to substantive questions relied on the social worker pointing at the anatomical doll's genitals and asking Girl A "What is this long thing?"; during the event investigation, leading questions were used, and when Girl A's answers did not fit the interviewers' expectations, questions were repeated, her answers ignored, and details she had never mentioned were added; and the gender equity investigation proceeded through leading combined with game-play, leaving Girl A unable to understand whether the interview's purpose was to obtain a true accusation or to play, and making true accusation and role-play even harder to tell apart.

When teacher Hsu Pei-ming was convicted, this body of forensic psychology research was not yet familiar to Taiwan's legal community. But now, it is time to re-examine this case through that lens!

Related Links

"Go Pei-ming" Facebook Page