The Christchurch Civic Creche Case

News Reports Index

2003 Sept



New Zealand Listener
September 13-19, 2003.
Vol 190 No 3305.
(published Sept 6, 2003)

Memory and the Law
Why the nature and accuracy of psychological evidence is on trial in the Peter Ellis case.
by Michael C Corballis
Professor of Psychology at the University of Auckland.

We need to change dramatically if psychology as a discipline is to lead the way in avoiding harm to all those who are affected by the consequences of both accurate and false recollections of abuse." - American Psychological Association

To take an example from Sas's report to Eichelbaum, if abuse causes nightmares, she suggests, then a child who has nightmares has been abused.


Most people who have taken the trouble to read Lynley Hood's book A City Possessed will surely have strong doubts as to whether Peter Ellis was guilty of any of the charges laid against him in the Christchurch Civic Creche affair. The book has been widely praised by reviewers (including myself) and won the Montana Medal for Non-fiction in 2002, and this year Hood was awarded a Doctorate of Literature by the University of Otago. Needless to say, there have been some critics. Hood's focus on Christchurch as providing the right atmosphere for a witch-hunt is perhaps a little unfair the site could probably just have easily been Auckland and some have leapt to the defence of their radical feminist colleagues who were claimed (and in some cases named) by Hood as having helped to create that atmosphere. The real problem lay not so much in an individual city or in individual people, as in a worldwide hysteria over the incidence and alleged excesses of sexual abuse, a hysteria that overwhelmed the legal system and, for a while, the population at large.

Nevertheless, the book is a compelling document, and stands in marked contrast to the report produced by Sir Thomas Eichelbaum, the judge who undertook the most recent appeal against Ellis's conviction. The Eichelbaum report concluded that the appeal failed "by a considerable margin". Part of the difference, of course, is that Eichelbaum was constrained by his terms of reference, whereas Hood enjoyed the freedom, and the time, to examine any aspect of the case as she saw fit. Even so, it is difficult to avoid the feeling that, had the report been undertaken by someone such as Hood, an independent citizen armed with common sense and a scientific background, the appeal would almost certainly have been upheld. Can the law really be such an ass?

Part of the blame, though, might fairly be levelled at the discipline of psychology, which had quite a strong hand in both the original trial and in the appeal. An editorial in the New Zealand Law Journal (February, 2002, p1) had this to say:

"Little psychological evidence stands up to serious scrutiny. Psychologists have managed to con the system for years with nonsense such as 'offender profiling' which has no scientific basis whatever. The fact is that psychology completely lacks a general theory of human behaviour and the divisions between schools of psychology are as deep as arguments about whether the Earth goes around the sun or vice versa. Few psychologists understand the logical structure of evidence they are giving, as surveys of numerous cases, listening to them speak on this and other issues at seminars, and personal experience of trying to train them in evidence-giving demonstrates. Almost no statements made by psychologists are backed up by the population data necessary to give the evidence probative value. The so-called 'prosecutor's fallacy' is endemic. Recently, a psychologist on television suggested that many premature births are due to stress events in pregnancy. To prove this she interviewed mothers who had given birth prematurely and discovered that some high proportion of them had suffered stress events in pregnancy. This, she said, proved her theory. Much psychological evidence in real court cases in New Zealand and elsewhere has been as unintelligent as this. The mystical hold that psychologists seem to have over the legal system should be broken."

Could it be, then, that psychology is the ass?

As the above extract makes clear, psychology is a deeply divided discipline, but that does not mean it has nothing to offer. It was established as a scientific discipline in the late 19th century, and until shortly after World War II was largely dominated by academic psychologists. After the war, however, there was a widespread need for improved services in mental health. A famous conference of psychologists held in 1949 in Boulder, Colorado, established the so-called "scientist-practitioner" model for the training of professional psychologists. But over the course of time, scientists and practitioners have gone their separate ways, and practitioners have come to greatly outnumber those responsible for developing the science of psychology. This has led to a serious erosion in the scientific standards exercised by many practising psychologists.

The rift is not absolute, of course. There are clinical psychologists, most of them working in academic or institutional settings, who have maintained rigorous standards of research and some mainstream academic researchers who have not. Moreover, some counsellors and therapists who have been instrumental in scaremongering over the prevalence of sexual abuse have been social workers, not psychologists, and the leading (one might also say misleading) professional in the Ellis case is a psychiatrist. One of the most influential publications in fomenting sexual-abuse hysteria was the book Courage to Heal, by Ellen Bass and Laura Davis, who have acknowledged that their recommendations were not based on scientific evidence. A good many pseudo-experts travelled the world in the 1980s and 1990s, holding workshops or public meetings, further intensifying the witchhunts that permeated that period. But the rift within psychology is real and serious, and has not been helpful to legal counsel.

Nowhere is that rift better illustrated than in the attempt by the American Psychological Association (APA), the world's largest psychological organisation, to establish guidelines on the nature and interpretation of memories of childhood abuse. In psychological terms, the accuracy of memory, and especially childhood memory, is pivotal to the Ellis case and to many similar cases that arose around the world in the 1980s and early 1990s. The working party to investigate these matters comprised three clinical psychologists and three psychological scientists whose research was directly concerned with memory and the reliability of childhood memories.

The two factions found themselves unable to agree, and eventually published their "final conclusions", a largely unhelpful document that sets out the relatively few points of agreement. The concluding sentences of this document, echoing the statement quoted above from the New Zealand Law Journal, run as follows:

"We are fast becoming a collection of psychologies, each uninformed by the data and epistemologies of the others; in short, we axe pluribus but not unum. Most critically, we need to change dramatically if psychology as a discipline is to lead the way in avoiding harm to all those who are affected by the consequences of both accurate and false recollections of abuse."

More interesting than the document itself, though, are the follow-up statements from each of the factions. Both refer to extensive psychological literature, with the clinical faction giving much more weight to claims of memory repression, post-traumatic stress disorder (PTSD), dissociation and the like, and the scientific faction, which included internationally known experimental psychologists Stephen J Ceci and Elizabeth F Loftus, expressing strong doubts about the reality of these concepts, and referring to the more basic literature on memory and child testimony. The document and the two statements are published in Volume 4 (1998), Number 4 of the journal Psychology, Public Policy and Law, and should be required reading for anyone seriously interested in the psychological issues of the nature of memory and its implications in legal cases over claimed sexual abuse.

The issues are difficult and complex. Memories refer to events in the past, and it is often difficult to verify exactly what happened. Sexual abuse, when it occurs, is likely to be furtive, so that there are unlikely to be third-party witnesses, and in cases of ritual abuse there may still be a conspiracy of silence. This means that research on memory may be contaminated by bias and political agendas, and it is important to maintain high standards of objectivity and critical analysis. Our understanding of how memory works has greatly improved even since the publication of the APA documents, and is based on converging evidence from experimental psychology and neuroscience. The surge of interest in the science of memory was indeed due largely to the controversies over the nature of memory for abuse, and the reliability of memories for childhood events. It is now widely recognised, at least among psychological scientists, that episodic memory (memory for specific events) is highly fallible, and experiments showing how easy it is to create strongly believed false memories are becoming standard undergraduate laboratory exercises.

Politics often strikes at the most vulnerable points of science, and leads to the use of intimidation tactics. This has been evident throughout the period of hysteria over sexual abuse, and also tainted the Ellis trial. Hood tells of "M", the clinical psychologist assigned to the defence in the Ellis case, who resigned as a result of pressure not to reveal information about the unreliability of child testimony. As quoted by Hood, M was "scared shitless". A few years ago, Loftus, probably the world's foremost expert on the fallibility of memory and a past-President of the American Psychological Society, was invited to address the annual meeting of the New Zealand Psychological Society. Loftus's 1994 book The Myth of Memory Repression is a strong statement against assumptions that are often used in clinical practice, especially in cases of sexual abuse, and has made her a frequent target of attack. In an appalling disregard for academic freedom not to mention courtesy a number of members of the society tried to prevent Loftus from speaking. To its credit, the society managed to defuse the situation and the address went ahead.

Nevertheless, the society is dominated by practitioners and remains somewhat vulnerable to polemics. Few mainstream academic psychologists belong to it, in part because its fee structure is oriented towards services to professionals rather than researchers. In our universities, there is nevertheless a strong tradition of basic psychological research, including research on memory and on child development, which are directly relevant to the Ellis case. This work seldom features at society meetings, and may not be widely known to the judiciary or to the public. Those interested in fundamental psychological science are more likely to attend more research-oriented conferences elsewhere. This is again indicative of the split between the scientists and the practitioners, and parallels what has happened in other countries. For example, in the US, some APA members felt that the association was overly dominated by professional matters, and split off to create the now flourishing American Psychological Society, which is primarily a scientific organisation. In Australia, an annual Experimental Psychology Conference has long replaced the professionally oriented Australian Psychological Society as the main meeting place for scientific psychology. In 2004, this conference will be held for the first time in New Zealand.

This confusing state of affairs might partly explain why Eichelbaum did not manage to obtain better scientific information concerning the nature of the evidence that was used to incriminate Ellis. Part of the problem lay in his choice of experts. His terms of reference instructed him to obtain opinions from "at least two internationally recognised experts". He chose exactly two. In deciding whom to choose, he consulted the parents of the complainants, and then requested the assistance of Legal Counsel within the Ministry of Justice. He "also had a long discussion with an [sic] USA Law professor who knew or was familiar with the work and reputation of many of those under consideration". He apparently did not seek the advice of any psychologists, even though there is considerable expertise in several departments of psychology within New Zealand on matters directly relevant to the enquiry. Eichelbaum also rejected two international experts who had been engaged earlier by Ellis's counsel, as well as a third who had provided input for a television programme on the case, on the grounds that if these experts were chosen, the report "would lack credibility in the eyes of the parents, the interviewers and the other participants concerned". Lack credibility? The first two experts were Stephen J Ceci and Maggie Bruck, probably the world's foremost authorities on the interviewing of children which is precisely why Ellis's counsel engaged them. One also wonders whether Ellis or his counsel were consulted in this matter. Did either of the experts lack credibility with Ellis?

In the event, one of the chosen experts was a Canadian psychologist, Dr Louise Sas, who had been involved in a number of cases of alleged sexual abuse in North America. Can she really be considered an expert? None of the journals she has published in is listed in Journal Citation Reports, a compendium of 5876 journals that provide the benchmarks for academic and scientific integrity. Sas has acted primarily as a policy adviser and advocate for victims of child abuse. She recently achieved notoriety in Canada, where she presented evidence of sexual abuse against a former nanny, who was eventually cleared of any wrongdoing. An editorial in the Toronto Globe and Mail on Saturday January 20, 2001, focused specifically on her role, and quotes a comment by one of the lawyers that "Dr Sas can interpret every fact and every behaviour as evidence of abuse".

This illustrates the well-known logical fallacy of affirming the consequent, which has long been endemic to the diagnosis of sexual abuse. To take an example from Sas's report to Eichelbaum, if abuse causes nightmares, she suggests, then a child who has nightmares has been abused. The absurdity of this is illustrated by the analogous argument: if someone is murdered, then that person is dead. A dead person, the argument goes, must therefore have been murdered. My understanding is that the interpretation of psychological symptoms as evidence of abuse is now thoroughly discredited and, in this and other respects, Sas's approach seems curiously dated.

In her report to Eichelbaum, Sas's choice of words often betrays her implicit assumption that Ellis was guilty, as indeed were others accused of sexual abuse. Children do not provide testimony, they "disclose", and in so doing do "the right thing". She accepts that some interviews with the children were faulty, but then refers to this as "unfortunate", and excusable because of the "interviewer's need to get a disclosure". A person accused of sexual abuse is an "offender". Cases where claims of sexual abuse are upheld by the court are described as "positive court outcomes". And so it goes wonderful material for anyone interested in discourse analysis. This is adversarial writing, not the dispassionate report one should expect of an expert called upon to advise in a legal case.

The other expert chosen by Eichelbaum was Professor Graham Davies, a British psychologist known for his involvement in helping define "good practice" in interviewing techniques. He does have a moderately respectable record of published research, although not in the major journals of memory or developmental psychology, and one can easily find in New Zealand psychologists with better research records on both basic memory processes and the reliability of testimony. Davies might be described as middle-of- the-road with respect to psychological disputes over the reliability of children's testimony. In keeping with this, his report on the Ellis case notes marked deficiencies in the interviewing techniques, but then excuses these by noting that "one can sympathise with the interviewers". And, having noted these deficiencies, he rather unaccountably concludes that "by today's standards the interviewing stands up surprisingly well". Davies's rather see-saw report clearly gave Eichelbaum the opportunity to conclude that the interviews were acceptable, and to find concordance between the two experts.

At one point, Davies notes that the allegations "need to be studied in the wider context of the investigation. For instance, do the toilet facilities at the creche correspond in their layout and construction to those described by the children?" Eichelbaum did not investigate the physical layout of the creche, because it lay outside the terms of his investigation. He writes:

"One of the grounds of the 1999 Appeal was the alleged non-disclosure of photographs of the interior of the creche, and [apparently] of creche activities. Having regard to the wording of the Terms of Reference ... neither the international experts nor I have been shown those photographs."

Yet other accounts, including Hood's, make it clear that many of the alleged charges against Ellis simply could not have taken place in the environment of the creche.

It is likely that the debates over the Ellis case will continue to pit clinical experience against objective science a classic issue in psychology (as well as in medicine). In some cases, science clearly wins. In a recent highly publicised case, for example, David Dougherty was jailed for rape, but DNA evidence subsequently showed another man to be guilty. This illustrates that some kinds of evidence are more or less irrefutable. Psychological science has not yet progressed to this level of irrefutability, although advances in brain imaging may eventually lead to better evidence, and help determine which memories are true and which are false. This is not the same as lie detection, since false memories are typically honestly held although, as any parent knows, children do sometimes lie.

If psychological expertise is to be invoked and I think it must be then it is imperative that it lean as far as possible towards the findings of psychological science. After all, science is all about evidence, and is subjected to rigorous standards of objectivity, accuracy and logic. It is sometimes wrong, of course, but it is subject to self-correction, and in any case is the best kind of evidence we've got. In her 1979 book Eyewitness Testimony, Loftus recommends that court experts on matters of testimony should have a higher degree in experimental psychology and a record of published research. Experience in clinical practice of itself will not do human interpretation of experience is notoriously prone to bias and superstition, and it is the scientists, not the clinicians, who are most likely to have the most comprehensive and up-to-date knowledge of the relevant areas of research.

This is not to denigrate the role and importance of clinical psychology, or of clinical training. In our universities, degree programmes in psychology are designed to instruct students in the science of human behaviour and cognition, just as medical programmes instruct in the biology of the human body. Nevertheless, students often resent having to learn statistics, experimental design or basic psychological principles, and there is sometimes the suggestion that they are selected into clinical programmes on the basis of societal concern, or of personality, rather than scientific aptitude. There is an ever-present danger that scientific aspects of training are diluted or overridden by political movements or outbreaks of hysteria, or that practitioners are given insufficient incentive to keep up with scientific developments after they have graduated.

It is sometimes suggested that if Ellis were to be found innocent of wrong-doing, then real offenders would also go free, and people would no longer believe what their children tell them. It is as though innocent people must be convicted in order to right a societal wrong, an argument that evokes the worst excesses of puritanism. This is nonsense. The aim of scientific inquiry is to establish when children's testimony can be believed and when it cannot, when memories are reliable and when they are not, and also precisely what constitutes harmful sexual activity and what does not. The goal must surely be to ensure accuracy, so that genuine sexual abuse is detected, but innocent people are not convicted. We may never achieve perfect accuracy, but science can surely take us closer to that goal. And it is possible, though I think unlikely, that the opinions of scientifically minded experts might have led to Ellis being found guilty.

The editorial from the New Zealand Law Journal, quoted earlier, concluded with the sentence, "The mystical hold that psychologists seem to have over the legal system should be broken." The main point of this article is that the law cannot really do without psychological expertise in cases involving psychological matters, such as memory-based claims concerning past events. The problem is not to eliminate psychological testimony, but rather to separate good psychology from bad psychology - or good science from bad science. This is something that universities do all the time, in hiring (and firing), in promotion and in awarding research grants. University academics know which journals publish reputable research and which do not. They know which of their national and international colleagues are the leaders in their fields. To be sure, academics sometimes disagree, but anyone who has served on a promotions or research committee knows that consensus can usually be found, especially where the evaluation of people is concerned. It seems extraordinary that no department of psychology in this country was consulted over the provision of expertise in the Eichelbaum appeal, or even, to my knowledge, in the original trial.

The Justice Minister has avowed that he will not order any further inquiry into the Ellis case unless new evidence is forthcoming. The problem lies not in the lack of evidence, but rather in the manner in which the existing evidence was interpreted. That message is clear from Hood's book, but somehow seems to have escaped the legal system.


[Some of the information in this article comes from an unpublished report by Jonathan Harper, whom I gratefully acknowledge.]