• The Listener
  • North & South
  • Noted
  • RNZ
Photo/Getty Images

Child testimony: Could the Peter Ellis case happen again?

Peter Ellis will forever be associated with one of the country’s most troubling child sexual abuse cases. But his experience has at least helped to bring changes to the way children’s evidence is gathered and used in court.

When author Lynley Hood interviewed Peter Ellis in jail while researching her book A City Possessed, on the Christchurch Civic Crèche case, she asked him for some biographical details to flesh out his character. Ellis raised one eyebrow and asked archly, “How graphical do you want it to be?”

Despite spending seven years in prison in the 1990s, Ellis’ spirit was undaunted, Hood says. It pointed to a character with more steel than many imagined, a trait further illustrated after his release in 2000 by his ceaseless efforts to clear his name. His supporters hope that despite Ellis’ death on September 4 from bladder cancer, aged 61, a final appeal to the Supreme Court in November will finally see him exonerated of the child sex abuse charges on which he was convicted in 1993.

Now, as obituaries are written rather than his biography – however “graphical” that might be – Hood and others are reflecting on the impact of his case and how it will be viewed in New Zealand’s judicial history. What change has it brought about? What is the legacy of Peter Hugh McGregor Ellis?

Read more: The Kiwi researcher trying to improve the reliability of child testimony | Why the Peter Ellis Civic Creche Case won’t go away

Hood’s book concluded Ellis was the innocent victim of a city in the grip of mass hysteria over satanic child sexual abuse – a city that needed to find a scapegoat. “Hundreds of high-profile criminal cases have gone from notoriety to obscurity in the 26 years since the Christchurch Civic Crèche case hit the headlines,” Hood told the Listener after his death, “but public disquiet over Peter Ellis’ convictions isn’t going away. The ripples spread too wide.” These included an exodus of male teachers from the profession. According to Ministry of Education figures, currently less than 5% of preschool teachers, and less than 12% of primary school teachers, are men.

Peter Ellis. Photo/John McCoombe/Listener

Nigel Hampton QC, Ellis’ former lawyer, believes that if a Criminal Cases Review Commission is established, the Ellis case will have been instrumental in its creation. “Our criminal appellate system is too constrained by statute and convention to even have before it, let alone try to reason with and decide upon, the wider background issues and context – including the epidemiology of the satanic ritual abuse moral panic – that played such a significant part in the Ellis case.”

For law professor Mark Henaghan, formerly of the University of Otago but now at the University of Auckland, the Ellis case was a watershed of “how best to obtain evidence from children and how not to do it”.

At the University of Otago, associate professor of psychology Rachel Zajac works closely with Harlene Hayne, the university’s vice-chancellor, whose research team’s forensic analysis of current thinking on children’s evidence in such cases is pivotal to the appeal. Hayne will not speak about the 1000 hours of work since her involvement began in 2004, but Zajac says the criminal justice system failed Ellis repeatedly.

“Although I’d like to think we’re finally on track to correct this, any correction now comes too late. It’s our responsibility to make sure we learn from this case and keep on learning. It will always be a reminder that when interviewing a child – or indeed any witness – good intentions don’t cut it. We need to be doing everything possible to elicit memory evidence in a way that preserves accuracy, and we need to ensure that memory evidence is presented and evaluated appropriately in the courtroom. To do that, we need to connect science with practice; we need to connect what we know with what we do. We all owe that to Peter.”

Lynley Hood. Photo/Supplied

Questions, questions

The flaws in the way the children’s evidence of abuse was gathered and presented in court included contamination by parents and “expert” interviewers who used leading questions, suggestions and rewards to elicit disclosures in multiple interviews. Zajac says the risks of repeated leading questions, and ascribing behavioural issues as symptoms of abuse, were known even then. “I don’t think we can say we didn’t know at the time. We certainly didn’t know as much, but we did know.”

Zajac has done several studies on the cross-examination of children, a practice that, she says, contravenes everything recommended for child interviews. “You’re allowed to lead the witness – in fact, you’re encouraged to lead the witness. The questions are often very complex, very confusing to the child, and they are quite confrontational because, by definition, cross-examination is trying to discredit them.”

Despite this, she says, courts must retain the ability to cross-examine. “We need to be able to test children’s evidence, especially in these cases where it’s all you’ve got. The vast majority of these hinge on the word of the child versus the adult. I don’t really have a solution for it, but I do know that the questioning style used during cross-examination is detrimental to children’s accuracy.”

She obtained transcripts of courtroom cross-examinations and found the children’s responses were “really concerning. Even when the questions were confusing and ambiguous, children were really reluctant to ask for clarification. We saw quite a bit of evidence of misunderstandings where children would answer a question that hadn’t actually been asked. And when leading questions were asked, children were highly likely to say ‘Yes,’ rather than ‘No, you’ve got that wrong.’”

Three-quarters of the children made at least one significant change to their earlier testimony. “If a child had said in an evidential interview that the colour of the car was red, when the lawyer said, ‘Oh, perhaps you’re mistaken, perhaps the car was blue,’ they started saying, ‘Okay, the car was blue.’ Some children retracted altogether and said, ‘Okay, you’re right, it didn’t happen.’ But of course, we know absolutely nothing about accuracy in those situations. It could be that it’s helping children to become more accurate – lots of advocates of cross-examination would say it clears up misunderstandings, but we thought, given the questioning style, it’s equally possible the children were starting off correct and ending up incorrect.”

Forensic analysis: Harlene Hayne, above, and Rachel Zajac. Photos/Supplied

To test this thesis, Zajac’s team recruited large groups of children – originally five- and six-year-olds but later including older children up to adolescents – and staged a novel event about which they were interviewed on video later. In one study, children were taken on a visit to a police station where they had their mugshots taken and went into a jail cell. “But there were also things they didn’t get to do that we asked them about later.”

When interviewed immediately afterwards, the children were highly accurate in their recollections. They were then reinterviewed eight months later – the sort of delay that occurs in court. “We showed the children their original video, and then interviewed them using a style consistent with what we see in court. We were trying to talk the child out of their earlier responses. So, if they said they had gone into the jail cell, we tried to get them to say that they didn’t. We had a number of standard reasons for disbelief that we had seen in the transcripts and they ranged from, ‘I think you’re mistaken about that,’ to, ‘You’re making that up,’ or, ‘Someone just told you to say that,’ or, ‘If your teacher said you did get to do that, she’d be right, though, wouldn’t she?’ And the children would often just go along with those things – they were just as likely to change a correct answer as they were to correct a mistake.”

So, why would children do this? Zajac says researchers have suggested many reasons, including that the delay made them forget what they did or didn’t do. But she says her team replicated the research and reinterviewed the children two days after the event and “we still saw the same thing”. About a third of the children changed their “evidence” about all four things they were questioned about, and only about 15% remained true to all their original responses. For most, she says, “their accuracy plummeted to the point where they were basically operating at chance, even those who’d started off at 100%”.

Which, of course, raises the question: why couldn’t the lawyers in the Ellis case cross-examine the children into recanting their evidence? But, at trial, the defence tactics focused on eliciting details of the contamination by parents and “expert” interviewers rather than encouraging the children to admit they could be mistaken. Given the number and extent of interviews they’d already been subjected to at such a young age, it was highly likely they were convinced the abuse had, in fact, happened.

Zajac says she doesn’t like people to refer to children’s evidence as “unreliable”. The onus is on the adults to get it right. “We know children have special vulnerabilities and it is up to us to interview them in a way that those aren’t played out. When you have a child who is not quite sure what is expected and wants to please an adult, and an adult with very good intentions but a strong belief that something has happened to this child, then, yes, you have a recipe for disaster.”

It’s difficult to counsel parents against sustained questioning of a child they fear might have been abused, she says. “You think of course those parents would have questioned those children within an inch of their lives, it’s a well-intentioned thing. But if the child mentions something, just say, ‘Tell me everything you can remember,’ thank them for telling you, then take it to the professionals.” She says most of the child interviews she sees these days as an expert witness are “really well conducted”.

Zajac is an associate investigator under Hayne on new research looking at how jurors interpret the evidence of both children and adults in cases of sexual assault, especially when they are historical, and what warnings or instructions might be useful.

Law professor Mark Henaghan. Photo/Alan Dove/Supplied

Where there’s smoke …

People who believe Ellis was guilty point to the consistency of the children’s stories about the extreme and bizarre nature of the abuse. Ellis was accused of sodomising children, forcing them to eat his faeces, urinating on them, suspending them in cages, putting them in ovens or taking them on terrifying trips of abuse through tunnels, ceilings and trapdoors. “I’ve always had the theory that these are the worst things that children can come up with and when a child is interviewed repeatedly with a strong suggestion that something bad happened here, when they finally go, ‘I’m going to give this person what they are asking for,’ they do tend to come up with remarkably similar things across cases, with the tunnels and cages and sacrifices and things like that. And when one child reports tunnels, interviewers ask other children about the tunnels and things go downhill from there.”

People, including jurors, might have a “where there’s smoke, there’s fire” approach to implausible allegations, Zajac believes. “They can say, ‘Well, of course that didn’t happen, but something must have happened for the child to be saying that.’”

Although the Ellis case will always be held up as pivotal to the later overhaul of approaches to child witness testimony here, there is evidence that, internationally, change was coming anyway – just not quickly enough for Ellis.

In 1993, the year after he was arrested, two of the world’s foremost experts in the field, Stephen Ceci of Cornell University, Ithaca, New York State, and Maggie Bruck of McGill University, Montreal, reviewed the suggestibility of child witnesses after a raft of sex-abuse allegations at day-care centres, the first of which happened at the McMartin Preschool in California in 1983, resulting in six years of trials but no convictions. In 1988, New Jersey day-care teacher Kelly Michaels was sentenced to 47 years’ jail on 115 counts of sexual abuse in a similar case. Parents wearing “Believe the Children” buttons had packed the courtroom to hear the evidence of children who claimed Michaels made them eat boiled babies, put swords into their rectums and played the piano naked. Michaels served five years in prison before she was exonerated in 1994.

Despite those “Believe the Children” buttons, Victoria University senior lecturer Deirdre Brown says we can’t always do that. “There are multiple reasons why a child might make a false report, including pressure from others, wanting to comply with what they think the interviewer wants, developing a false memory – which feels real, and is not necessarily an intentional deception – being questioned inappropriately and repeatedly, concern for consequences and so forth. We should also not dismiss the capacity of children. We know they can provide useful descriptions of their experiences when interviewed well. So, should we believe the children? It depends.”.

Exonerated: New Jersey teacher Kelly Michaels. Photo/Getty Images

Asking the right questions

Despite the lessons learnt in the Ellis case, the potential remains for “moral panics”, witch-hunts, and contamination of children’s evidence by parents or well-meaning but unskilled interviewers, lawyers say. For example, staff at Oranga Tamariki (the Ministry for Children) question children before formal evidential interviews, but exactly what is asked, and how, is often not recorded. However, researchers’ concerns about the friction between court processes and the accuracy of child testimony are being taken seriously in legal circles, and there are efforts to make practical changes.

Whangārei lawyer and academic Emily Henderson, whose research has focused on trial reform, says cross-examination of vulnerable and child witnesses “is pretty much the perfect storm”, because it is so easy to get children to contradict themselves or to look foolish, contradictory and unsure. “We’re now sitting on 30-plus years of research that demonstrates over and over again in all the comparable jurisdictions that conventional cross-examination of vulnerable people is shooting fish in a barrel.”

She says as well as lawyers being fairly laissez-faire about whether the witness meant to retract or not, adapting questioning to the language levels of children or vulnerable adult witnesses is much harder than lawyers generally realise. “Much of what has been found is the result of simple linguistic incompetence on the lawyer’s part rather than deliberate manipulation.”

Henderson, who completed a PhD at Cambridge University in 2000, later looked at changes in the UK that had improved the experience of those witnesses. From about 2010, “judges suddenly started taking control of their courtrooms, first by writing appeal court judgments that made cross-examiners comply with what the research was saying … and stopping some of the more obvious and egregious forms of cross-examination.”

When she returned, she and Whangārei judge Duncan Harvey launched the Whangārei Child Witness Project, in 2014, which cut waiting times at court before children gave evidence and asked lawyers to rephrase leading or complicated questions. “After the initial shock, lawyers embraced it – they could see why these things were being done and were happy to go with it and didn’t think it was unfair to defendants.”

Judges started looking for “tag” questions, where lawyers made a statement and then added a tag such as “didn’t he?”, “didn’t you?”, “wasn’t it?” or, “So, daddy didn’t do it, did he?” Also targeted were multi-part questions. “So, when you got out of the car and you saw the red car coming towards you and the three men got out of it, did you notice the gun?”

“The poor old witness might answer, ‘No, I saw the gun later,’ but what they’ve been taken to agree to is three men, a red car, that they got out of the car first … all of that.”

Pilot sexual-violence courts in Whangārei and Auckland that used the approach will now become permanent.

Ellis’ former lawyer Nigel Hampton QC. Photo/Getty Images

Liberal homophobia?

Hampton believes his former client’s flamboyance and homosexuality played “a considerable part” in the case. “Despite, on the surface, the liberalness of the parents saying we would love our children to be exposed to something different, within them there was still, I think, considerable underlying conservatism about such things as gayness and it didn’t take much to trigger that when the suggestion arose. I’m playing pseudo-psychologist here when I’m not qualified, but that’s my take on it.”

Ellis’ current lawyer, Rob Harrison, declined to comment on the case itself, but says he spent time with Ellis days before he died at the Nurse Maude Hospice in Christchurch. He says Ellis spoke of the support of locals where he lived in Leithfield. “Here he is facing imminent death and what he’s talking about is how grateful he is to the community that took him in, in North Canterbury. I think that tells you a lot about him.”

Harrison has applied to the Supreme Court for the appeal to still be heard, despite Ellis’ death, after the court earlier indicated it would be open to doing so. The hearing is scheduled to start on November 11.


Cardiologist Mark Webster. Photo/Adrian Malloch/Listener

An epidemic of the bizarre

Allegations of unusual mass sexual abuse against children peaked in the US between 1981 and 1995 – then stopped.

Support for efforts to clear Peter Ellis’ name came from many quarters. One unlikely figure who put in many hours of unpaid research is Auckland consultant cardiologist Mark Webster. Webster combed newspaper databases from OECD countries to chart the history of cases similar to that of Ellis, in which there was no physical evidence of abuse, more than one person was accused and, when unusual or bizarre aspects were described, there was no physical evidence to support them. Using the criteria of sexual abuse against children aged between two and 10, occurring in a crèche, childcare, preschool, kindy, Sunday school or foster home and involving more than one victim, Webster found 36 cases. In the US, these cases occurred between 1981 and 1995, then stopped. In the rest of the world, they spanned from 1986 to 2010.

“The pattern shows some similarity to an infectious disease epidemic, starting in the US and then spreading to the rest of the world. The outbreak was almost over in the US by the time Peter Ellis was charged.”

Webster says after reading A City Possessed in 2001, when it was first published, he’d thought an inquiry would be launched that would “sort it”. “But nothing happened.” He has sent his findings to Ellis’ lawyers.

The early and high-profile cases in the US were “strikingly similar” to Ellis’, says Webster, with many children of similar age allegedly abused and bizarre allegations. “It seemed to start from nothing, and then [the cases] just disappeared. And you’d have to think, why have they suddenly disappeared?”

Accusations against Peter Ellis in 1992 came as a spike in similar cases in the US was declining but allegations in other OECD countries were rising.

In New Zealand, another case of alleged sexual abuse at a preschool – albeit without the bizarre allegations of the Ellis case – flew largely under the public radar about the same time. In 1994, a year after Ellis was jailed, Wellington Hospital crèche worker Geoffrey David Scott was convicted on eight of 20 charges and jailed for seven years. As with Ellis, the allegations were made in 1991 and 1992. And, as in Ellis’ case, Scott’s lawyer didn’t believe he did it. Pat Grace, who is now a Family Court judge, told the Listener it has always stuck in his mind that, with one child witness in particular, the allegations came out only after several separate interviews. “The first time, the child said nothing happened. Went into therapy and was re-interviewed. Nothing had happened. Put into therapy again. Nothing had happened. More therapy. Then interviewed the fourth time and, ‘Oh, yes, he did A, B and C to me. That would never get off the ground today, I believe.”

Grace says Wellington police contacted Christchurch police during their investigation. In one email, Christ-
church police urged their colleagues in Wellington to “think big”. “I thought it was bizarre. I think they clearly had formed a view that they were on to something and away they went.”

At the trial, he says, one of the jurors had been involved in “the sex abuse industry”. “We ran out of objections so ended up with someone I didn’t really want.” In the Ellis case, the guilty verdicts involved seven children, and five of their parents had worked in the sexual abuse field.

In both the Ellis and the Scott trials, the Crown’s star witness was psychiatrist Karen Zelas.

In another sex abuse case, the Court of Appeal, in a 2003 judgment, found Zelas to have “gratuitously” exceeded the scope of permissible expert opinion and could have contributed to a miscarriage of justice.

The Listener asked Zelas, who has now retired and writes poetry, to comment on her role in the Ellis case. She declined. “I have always had a policy of not speaking publicly about cases I have been involved in.”

This article was first published in the September 21, 2019 issue of the New Zealand Listener.