Few cases will ever compare to the three-decade saga of Peter Ellis. Lurid allegations in the early 1990s of bizarre child abuse rituals at a crèche in Christchurch produced convictions that stand alone in the level of public and judicial attention they have generated since.

The case has troubled the appellate courts multiple times and the Supreme Court has just delivered the last word, sweeping away all the judgments that preceded it in a unanimous decision and declaring the verdicts unsafe. Ellis, for his part, has gone to his grave without knowing vindication but the difficult and colourful case that defined his life has finally come to its conclusion.

The judgment turns, like all the previous appeals, on whether the expert evidence was properly given and indicates, if nothing else, how difficult the marriage of legal principle and complex science can be. The same issues with the same appellant have been traversed before with very different conclusions. Just how delicately the duties of an expert can be filleted, and where crucial lines of witness overreach exist, can be disputed by the highest-placed legal minds.

Multiple young children made multiple allegations when interrogated by professional interviewers and of central importance was whether they were believable. Much of what the interviewers got out of them was bonkers: anal assaults with needles and burning paper, caged children hanging from the roof and descent into dark netherworlds through trapdoors. Much more was simply indecent touching.

On the basis of the more conventional allegations, with minimal independent corroboration, Ellis went to trial on 28 charges relating to 13 children, none of whom was older than five at the relevant time (four female colleagues were also charged but these charges did not proceed). The issue for the Crown was how evidence from the mouths of babes would be received and it turned to its star witness, child psychiatrist Dr Karen Zelas.

Zelas wore multiple hats for the investigation, which gave the Supreme Court pause about her impartiality. She reviewed transcripts, supervised some of the interviewers and gave the police guidance. She also attended a meeting called by concerned parents and gave a television interview. At trial, the Crown relied on her, among other things, to outline the behavioural symptoms that indicate child abuse.

Evidential problems 

The central consideration for the appeal was s 23G(2) (c), as it then was, of the previous Evidence Act, which allowed experts to give evidence of childhood behaviour either “consistent” or “inconsistent” with being sexually abused. It had a long history of being difficult to apply and has no current equivalent on the statute books.

The problems were two-fold: it assumed there was solid science in this area, when all that experts could ascertain was behaviour that might or might not merit investigation for being abnormal. In other words red flags, but not proof. It also invited experts to flirt with the line of commenting on the credibility of the witness, which is not their territory. The section was supposed to allow opinion evidence about cohorts of like-age children formally studied, not the child before the court, but in practice the distinction was difficult to stick to.

Zelas outlined a range of suggestive behaviours, including nightmares, tantrums, fear of animals and obsession with penises. Although she acknowledged other stressors could cause the same thing, she said the symptoms were particularly suggestive of sexual trauma when they clustered.

In the Supreme Court’s view, this was the classic issue with s 23G(2)(c), which allowed evidence that confused symptoms suggestive of a particular stressor with proof of that stressor. All the symptoms were capable of, including when they occurred in clusters, was pointing to an issue to be investigated, but the jury was invited to understand them as determinative.

Zelas also opined at length on the memory capacity of the children by referring to specific complainants, which led the jury from understanding what could be expected of a certain age group, which would have been admissible, to a credibility boost for key Crown witnesses, which was not. She also engaged in reasoning the Supreme Court rejected for being circular: that the red-flag symptoms were more likely to indicate abuse because an allegation of abuse had been made.

All of this was held to be critical to the Crown, and therefore to the verdicts. The Supreme Court was careful to point out that Zelas did not have primary responsibility for knowing where the boundary of permissible opinion was, but she was permitted to so wander freely over the wrong territory that the verdicts could not stand.

Contamination risk 

The second issue was contamination. When news of the issue first spread, parents interrogated their children and formed networks that shared opinions, including a public meeting. Complainants were taken to therapists and evidential interviews were conducted multiple times on the same children.

Whether the resulting disclosures were reliable was a central feature of pre-trial submissions, and formed part of the reasoning that saw three of the co-accused discharged (the fourth was discharged when the single complainant citing her was not available). The same application failed for Ellis and much the same arguments rejected then were before the Supreme Court.

On the stand Zelas played down the risk of contamination and rejected the crucial proposition that young children do not have adult capacity to distinguish whether something in their recollection was real or suggested to them. A defence expert, Dr Le Page, said the opposite but with little elaboration.

The Supreme Court received fresh evidence supporting Le Page’s opinion, with the vital amplification that it is very difficult to rescue a contaminated statement with cross-examination. Even allowing for developments in the science since 1993, the Supreme Court held that what Zelas said about memory, contamination and suggestibility were contrary to known science at the time she said it. The Court of Appeal has been over this ground twice, in 1994 and 1999, and considered the jury heard adequate rebuttal of Zelas’s opinions to cure their flaws.

The Supreme Court held the rebuttal was too little and inadequately focussed. The trial judge told the jury that Zelas was the more experienced of the two. There was, according to the Supreme Court, a “substantial risk” that the jury deliberated with “a false sense of reassurance” on the risk of contamination.

So in the end the Crown has not proven any sexual abuse happened at Christchurch Civic Creche and the page can finally be turned. What remains to be answered is why it took three decades to get there. ■

Roderick Mulgan is a criminal defence lawyer and medical doctor with particular interest in medico-legal issues ■