Section Five

A Scientific Analysis of Other Factors Affecting the Reliability of the Evidence


(5) A Scientific analysis of other factors affecting the reliability of the evidence

(5.1) Memory distortion and memory construction through contamination
. (i) The Evidence for Contamination by parents
. (ii) Contamination by Children
. (iii) Contamination by Police
. (iv) Contamination by Child Therapists
. (v) Contamination by Adult Counsellors
. (vi) Contamination by Social workers
(5.2) The significance of delay between the alleged events and the allegations
(5.3) The significance of the ages of the children
. (i) Infantile amnesia and memory recall in young children
. (ii) Spontaneous false allegations in young children
. (iii) Children in the courtroom
(5.4) Suggestibility in young children
(5.5) Do the children’s "symptoms" of abuse have any scientific validity?
(5.6) Is consistency proof of reliability?
(5.7) Coercion
(5.8) A statistical analysis of the reliability of the experts’ evaluation of the evidence
(5.9) The effect on reliability of the proportion of impossible and unlikely allegations
(5.10) The retracted evidence
(5.11) Repressed Memory and Child Sexual Abuse Accommodation Syndrome
(5.12) Comparison with overseas cases
(5.13) Conclusions

Next Section
Previous section
Document Contents | Intro | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | References | Appendices |

(5) A scientific analysis of other factors affecting reliability of the testimony

"How dare they suggest such a thing."

Comment by Parent Ms X/Dogwood in regard to the suggestion made by interviewer Sue Sidey and Police, that it was possible that some of her child’s allegations arose from his efforts to please her.

The preceding two sections have already included a scientific analysis of the factors within the formal interview situation that have been shown to influence the reliability of the children’s testimony. The absence of corroborating evidence of abuse in this case means that the children’s testimony is crucial. In this case the alleged abuse occurred well before the formal interviews, therefore what happened during that intervening period must be examined. There was ample opportunity for the children’s memories to be contaminated by input from other people. The effect of this delay itself on memory accuracy needs to be examined as well as the intrinsic reliability of the memory processes of children of this age, and in particular, of the six children whose testimony resulted in convictions. In this case there was a major retraction by the most reliable witness and this also needs to be examined in order to compare its apparent original veracity (to the jury) with the testimony of the other complainants.

By way of introduction, I quote from a current university textbook on memory (Neath 1998 p326),

"... memories change or are constructed, usually without any awareness of the distortion…what is most impressive – or most disturbing – is how pervasive these changes can be."

There have been numerous cases of older children and adults inadvertently providing evidence from memory where central details such as the identity of an offender were incorrect. Such cases illustrate the need for caution. Loftus (1979) provides many examples from American cases.

A recent New Zealand case has direct parallels with the Ellis case in that the uncorroborated word of a complainant was given undue credence. In this case, a female student at Waikato University falsely accused a male student of raping her. He was charged, but the case fell through when it was discovered that the young woman had deliberately made it up.

Undue credence being given to uncorroborated testimony appears more common in sexual cases (where juries may become emotionally involved, and by their nature such crimes can be surreptitious).In another case in Santa Ana (California, America), Kevin Lee Green’s wife was badly beaten, and raped. Initially she could not remember the attack, but later (possibly after counselling and/or therapy) she accused her husband. After 17 years in jail, Green was released after another man confessed, and new DNA technology proved the other man’s claim. The judge publicly apologised to Green.

The 1993 New Zealand case where Nicholas Reekie was eventually convicted of the (1992) rape of an 11 year-old girl also appears to have been a case of mistaken identity. The girl in her statement the next morning thought she had recognised the voice of a neighbour (Dougherty) who was known to be violent towards his partner. From an examination of the girl’s statement, it is possible to see how she may have been genuinely mistaken. She stated that she was also blindfolded but that it slipped down. That night the full moon shone brightly as she was taken through the hole in the fence along the street from her home to three business areas where several bright floodlights provided very strong illumination. She stated at this time that the rapist had short hair (as did Dougherty, whereas Reekie had long hair), and that she positively recognised him as Dougherty.  

The rapist also replied to her question, "Are you David?" that he was. (Reekie had lived with Dougherty at one stage before the offence, so knew him). It is also possible that once that story was accepted (Dougherty was originally convicted of the rape) it became reinforced in the girl’s mind as a vivid and unshakeably genuine, but possibly inaccurate memory. To make matters worse for her, one police officer, Mark Franklin, and the media (who seem to have added a spin to the story- eg Donna Chisholm in Sunday Star Times 25 October, 1998) suggested that she was making up the rape allegation. This occurred immediately following a police re-investigation into the case, six years after the rape and after Dougherty's convictions had been quashed (it was another four years later that Reekie was charged). There has since been an apology from the police. Understandably, at the end of all this, her parents were upset at this claim, and do not feel sure, despite the new DNA evidence, and the many trials and appeals, of exactly what happened that night. Perhaps the possibility of two assailants remains open. The courts did eventually decide that the new DNA test and the girl’s testimony only indicated the presence of one rapist, and so the court ruled a two rapist scenario out. Dougherty was exonerated, and paid compensation.

The relevant conclusion that can be drawn from this episode is that an eleven-year-old had been sexually assaulted and raped, and yet still appears to have been inaccurate in her recall of some of the central detail of the incident. In the absence of clear hard evidence such as DNA profiling, it may be impossible in otherwise similar cases to be sure which aspects of the evidence are most reliable. The problem may be magnified in cases involving pre-school children.

Back to section contents

(5.1)     Memory distortion and memory construction through contamination

Before the formal interviews took place, there were many factors that could have generated false allegations from the children. If the children were exposed to worried parents, other children (who may have been developing mistaken memories), therapists, police and social workers (who strongly suspected abuse had occurred) then this exposure could, over time, be said to present a significant risk of contamination to the children’s memories.

Neath (1998) discusses how "memories change because they are continually constructed and reconstructed." Memories are not like tape recordings. In Neath’s words, "the generally reconstructive nature of recollection has been consistently supported" (by scientific evidence). So, there is ample scope, especially with young children, for extraneous (as well as internal) contamination.

Loftus (1993) discusses how post event information can become incorporated into memory, and shows how it is possible for people to construct entire events that never happened. Her first example is an anecdote from Piaget, who had a childhood memory of a non-existent kidnapping that was fabricated by his nanny in order to receive a reward. Loftus then discusses a paradigm of creating false memories of being woken at night by loud noises, "used extensively by other researchers who readily replicate the basic findings." Loftus writes of another example of research that revealed that some people misremember that they have voted in elections when they have not. Haugaard et al’s 1991 study is quoted as showing that people can be led to believe they have witnessed an assault, when they did not. More stunning is a study quoted by Loftus where there had been a sniper at a school (Pynoos and Nader 1989). Some children who attended the school had detailed memories of the event, yet they were nowhere near the events at the time. Loftus next discusses her own study where a false memory of being lost in a shopping mall was implanted. Finally, in the case of an adult (Ingram) arrested for child abuse, Loftus shows that certain adults are susceptible to believing they have committed crimes that they have not. All this research points to the strong possibility, that must increase over time, that false memories of entire dramatic events can indeed be constructed through talking to others.

Memories can be modified or distorted by the use of schemas (organised knowledge structures). An example of how our schemas can distort our memories in everyday life is found in French and Richards (1993) study. Adult subjects were instructed to look carefully at a clock face with roman numerals, and then draw it in detail from memory. Most subjects wrote the IV they expected instead of IIII that is used on clocks for the numeral four. A possible example of a schema at work could be the children’s unlikely descriptions of Ellis’s many accomplices. They seem to have been told (often by those they trusted and/or authority figures) that Ellis "is a bad man" who should "go to jail." It is possible their schema of criminals was one in which "baddies" act together in gangs. This effect may have influenced some aspects of the children’s multi-offender allegations.

Adults can confabulate (add additional information about things that did not occur – see Neath p328). There is no currently known evidence to reason that children could not also exercise this ability.

Attribution errors can be common in adults. An example, Neath (1998), involved a woman who had been raped whilst watching TV. She misattributed the face shown on the TV at the time of the rape to that of the rapist. Young children must also be capable of such errors. Neath states (p337) that

"Researchers in this area have concluded that there is no way to assess whether a memory is accurate or not without objective corroborative evidence."

As the work of many scientists (eg Ceci et at 1977) indicates, false beliefs can be sincerely held and resistant to change. In the Ellis case they may have easily survived the formal interviews, even when challenged, especially so given that the interviewers had deliberately tried to create a friendly atmosphere (conducive to "disclosure.") Only rarely, were the more credible allegations challenged, as noted by Ceci in section (3). Neath (1998 p 335) summarises that research confirms that subjects reporting misleading information are as confident and as quick to reply as were the subjects that reported accurate memories. Even offering the subjects a financial incentive will not improve accuracy of recall.

A recent study on parental questioning of children, (Bruck et al 1999) clearly reveals the dangers of repeated prior questioning by parents:

"Importantly, they [parents] had difficulty recalling how the information was elicited from their child, whether their child's statements were spontaneous or prompted, or whether specific utterances were spoken by themselves or by their child."

So there can be little doubt that construction of children’s inaccurate "memories "of abuse which never happened was highly likely in this case. Furthermore, such inaccurate memories are extremely difficult to discriminate from accurate memories of abuse that did occur.

These considerations are supported by a number of professional organisations:

"Research has shown that over time memory for events can be changed or reinterpreted in such a way as to make the memory more consistent with the person's present knowledge and/or expectations."

American Psychological Association, 1995.

"Memories also can be significantly influenced by a trusted person."

American Psychiatric Association, 1994.

"The AMA considers recovered memories of childhood sexual abuse to be of uncertain authenticity, which should be subject to external verification."

American Medical Association, 1994.

Back to section contents

(i)         The Evidence for Contamination by parents

The children in court, under cross-examination, revealed indisputable examples of parental contamination. For example, in court one child volunteered for the prosecution

"... some other parents rang up my mum and told her about that [the allegation] and she told me".

Examples also appear in the interviews themselves. The interviews of the children by parents occurred before the interviews that were conducted by professional child abuse workers. The initial allegations arose from these parental interviews. These interviews should therefore be logical targets for comparison with "best practice" interview techniques (as already discussed). It is relevant in this regard that Ceci et al’s 1987 study (quoted in Ceci and Bruck 1993 p 427) indicates that suggestions from parents may have more influence than suggestions from other young children in producing erroneous statements from young children. Once a false belief is well established (and there was up to 18 months for this to take place) any later "evidential" interviews, no matter how well conducted, merely serve as a vehicle for recording what the pre-schoolers have come to believe over that time. This is true regardless of whether the memories recalled are of real events, constructed events or a mixture of both. The NSW report (Eichelbaum p37) states,

"Under no circumstances should children be informed of other children’s complaints or involvement."

In the Ellis case the horses had well and truly bolted in this regard, well before the first formal videotaped interview.

Parents admitted using leading questions in the Ellis case (eg Sas p45). Sas states (p19) that:

"some of the parents talked with their children about the events as much if not more than the trained interviewers."

"There are some examples in interviews where the interviewer repeated the specific allegations they had heard from the parents…This tended to happen when children were not forthcoming with their disclosures."

In the case of Z/Kari Lacebark, Van Beynen (1995) states,

"[Kari’s] mother …admitted asking her daughter a number of suggestive and leading questions soon after the possibility of abuse at the crèche was first raised. [Kari’s] first allegation matched a question put by her mother seven days earlier. Yet not a single ‘how come’ question was asked by the interviewer".

Davies report highlights a case of parental contamination even though Davies himself failed to recognise this instance. It has already been noted in section (1.7.i) on bias, that Child Q/Lara Palm made comments in her first interview that appear rehearsed. According to Davies, Child Q/Lara also alleged during the interview that abuse occurred, at her own estimate, when she was aged "three or four" (p15 Davies). Davies describes that the child then went on to rationalise,

"...that Peter threatened that she would turn into a gherkin if she told anyone; she told her parents, she says, when there was no sign that this had happened by age 5 ("I realised he was not telling the truth")."

These are not the thought processes of a five-year-old. According to research cited by Davies himself (p3), in the case of 5 and 6 year olds "Information about when events occurred (day month, year,) is likely to be very poorly retained...". Moreover, he continues

"Elapsed time is generally poorly represented: children of this age have difficulty in estimating the duration of an event and how long ago it occurred."

Therefore Q/Lara Palm’s rationalisation in regard to appropriate time lapses, in order to justify alternative strategies to her continued silence and the exercise of logic in such a manner strongly suggests adult input. Yet Davies concludes, quite inexplicably,

"I see nothing in this interview that was redolent of coaching." (Davies p16).

The common statement that "nasty things happened" is also an indication of an adult narrative source.

Goodyear-Smith (1993b) significantly describes how one mother admitted in court how she cuddled her son and praised him for being brave each time he "disclosed" some more. Another girl, she notes, said in court that she had learnt her story

"before I went on screen" and that the interviewer "taught me what Peter did."

Another obvious source of parental contamination came through the use of books. Child Q/Lara Palm, for example (and others) was read two books by her mother prior to the first interview, What’s wrong with bottoms and A very touching Book. Her mother had also spoken to Q/Lara "no more than four or five times" about Ellis and the Civic childcare centre prior to the first formal interview. The often-repeated word, "yucky" uttered by most children would seem to have its source in a book, Katie’s Yucky Problem read to some childcare centre children by their parents.

The use of the two books that parents read to their children is significant. Sas and Zelas talked about children having "sexual knowledge beyond their years" (p50) which could "indicate" or was "consistent with" gaining that knowledge from an abuser (and note no other source was considered). Terms used by some children appear suspiciously identical to those found in these two books. But Sas blithely discounts this possibility as "not a concern" (Sas p45).

In his comments regarding contamination (p38,39) Davies accepts that there were no spontaneous allegations, no allegations arising from the first formal interviews, and that there was widespread questioning of children by their parents. Then he states

"Of more concern [than leading questions] is cross-talk between families against a background of persistent accusations against a suspect [Ellis]".

He makes an exception in regard to accusations set in the childcare centre toilets. In these he regards the crosstalk alone as insufficient to cause the similarities contained within the allegations. However he includes the qualification that (my emphasis)

"[the] degree of detail in some of these accusations [set in childcare centre toilets] is not in itself proof of the charges against Ellis, but ....[means] such accusations deserve to be taken seriously. They need to studied in the wider context of the investigation."

It can be inferred from his comments that Davies has little faith in any accusations other than those that involved the Civic childcare centre toilets, and even then this faith was qualified.

"Believe the children" was a highly visible child protection movement (CPM) slogan of the period (see below 3rd Parent meeting at Knox Hall). Often the slogan was coupled with the instruction to parents to praise a child for his/her "bravery" when having made a "disclosure". This is fraught with risk. It can result in unreliable testimony through a variety of mechanisms including; reinforcement of certain behaviours and through the failure to ascertain alternative explanations because of a lack of proper source monitoring. These have been already discussed in section (3). Although some CPM professionals during the Ellis case may have been aware of risks in this regard the damage had already been done. The slogan had been very effectively promoted to the public and little could be done retrospectively to temper its effects. This slogan seemed to acquire the status of a prime directive with some of the complainant parents, evidence for this is found throughout complainant parent Joy Bander’s book, for example (p38)

"We told him ["Tommy"/child X/Bart Dogwood] many times how much we loved him, how brave he was for telling us all those things, and how he was now safe from Peter and would never be hurt like that again."

This parent had even instructed the child’s siblings to respond in the same manner whenever the child spoke of Peter Ellis. Yet this parent would not accept her child’s initial statements that Ellis had not interfered with him and was in fact "his friend". Instead she chose to question him relentlessly over a period of three months (ibid):

"I decided to sound Tommy out from time to time .... and so about once a week I would introduce the subject."(p31)

"I couldn’t do that [let the subject rest after the child’s first formal interview] and so from time to time I began again to speak to Tommy about the Crèche and Peter Ellis as I had before. Once again I got nowhere. Each time I asked, I got the same answer: "no, there is nothing more to tell."" (p33)

[After three months] "I kept on trying to sound him out about the Crèche as before, but got nowhere." (p34)

[After four evidential interviews] "I felt he needed to be helped to talk some more, so I tried to draw him out further."(p50)

The case of X/Bart Dogwood as described by his mother is very illustrative of the contamination process carried out in varying degrees by some parents. It is therefore informative to examine the process of questioning outlined here in more detail.

His mother reports that she had talked to him since he was "quite little" about bad touching, and that he had been shown books on the subject (Bander p31). His parents first attended the third meeting of parents At least fifteen months had elapsed since the period when any of the incidents could have possibly occurred. Ms Dogwood then asked her son "a direct question which was, ‘did Peter touch his penis or bottom?’"(Bander p30). She then informed Bart about the meeting, and asked him questions over the next several days and then continued "about once a week…for a few weeks". "At first he said no [he had nothing to tell and...] that Peter was his friend." (Bander p30). Eventually he said something to his older brother who had been previously informed as to the nature of the inquiries

"he smacked my bottom really, really hard. I couldn’t hear the smack, but it really hurt.".

Ms Dogwood selectively reinforced these "disclosures" (as previously quoted)

"We told him ["Tommy"/child X/Bart Dogwood] many times how much we loved him, how brave he was for telling us all those things, and how he was now safe from Peter and would never be hurt like that again." (p38)

At the first evidential interview, a few days later, Bart began by saying (Davies p19)

"He [Ellis] was alright when I did go there…and now he’s not."

The interviewer then by degrees coaxes Bart to agree that Ellis "wobbled" his penis whilst changing him. He then repeated an allegation made by his parents involving another child, but "seems to have no personal experience of this" (Davies), stating that he knows

"cos mum and dad told me."

Parsonson (1999) states that the interviewers are giving the children permission (not normally granted by adults) to talk about scatological and "rude" things that they may naturally enjoy talking about. It is therefore unsurprising to find that a week or so later in the middle of a dinner party Bart said, to the horror of his parents, "You know, this carrot is like a big fat penis in your mouth to suck" (Bander p12). A feedback loop between the child’s home environment and the interviewer is now becoming established.

By the fifth interview, Bart talked of entering a secret room with the number 20 on the door. This room was accessed from the childcare centre via a ladder, and secret tunnel. In the room, Peter’s mother hung five children in individual cages from hooks in the roof. Various childcare centre staff stuck sticks and burning paper up their "bums." When challenged by the interviewer, Bart insisted,

"they really happened".

Davies (p23) points out in relation to this "repeated requests to recall may lead to construction of non-factual accounts, with the young child unable to distinguish between their retrieved memories, self-generated imagery and content derived from other, later sources. In X’s [Bart’s] case, it emerged at trial that all the video sessions were preceded by extensive discussions with his parents…interestingly, they [secret passages and cages] also figured in the testimony of children in the [now extremely dubious American] McMartin case. " The convictions were overturned in the McMartin case.

Ms Dogwood reveals the likely thinking behind her approach to the questioning of her son,

"I wish to explain what I have since learnt through my research on satanic ritual abuse, and the way children disclose abuse…Tommy [Bart] was scared. That is true and an understatement. A child disclosing about ritual abuse is an incremental and slowly progressive process. It may take a year or more before the entire story is told." (p85)

…"Tommy relied on my questioning, and in fact would say, ‘ask me a question about (this or that).’" (p86)

While Ms Dogwood does state that some of this rationalisation for her attitude and approach arose after the trial, the record shows that she began to adopt such an approach around the time of the first allegation of smacking made to Bart’s brother. Ms Dogwood quotes from Hudson (1991) on page 86, indicating that Hudson was another highly questionable source of [mis]information..

It may have been easy for Ms Dogwood to develop unreal and untested ideas about Ellis, as she

"had never had any dealings with him, nor any feeling about him, I began to wonder about him." (Bander p31)

Perhaps, had she met Ellis, she may not have developed such a harsh attitude,

in my opinion [re the Ellis sentence] it was a pity the death penalty wasn’t in." (p103 ibid)

Some parents brought along booklets they had made with their child to the interviews. These were discussed with the interviewers. Sas writes (p48), "…[it] was not something that I had seen before, but at least it gave the interviewer an opportunity to review…" Although this is permissible to Sas, it is not acceptable practice now (see section (4) Changes in Best Practice).

Hood (p443) reproduces information that one mother asked Eade to look into incidents of children being buried over their heads in the Civic childcare centre sandpit, with straws to breathe (physically impossible as the pit was too shallow), and being buried alive in coffins. This incident (and others) suggests that some parents had entered an emotional state in which they were becoming credulous.

Motivators toward behaviours resulting in contamination by parents may have been provided by two existing social support systems; social workers and the Accident Compensation Corporation (ACC). A parent support group was set up according to Ms Dogwood (Bander 1997) "sometime [prior to] August 1992). "It met either weekly or fortnightly, and was facilitated by a social worker." Early on in the case social workers provided a list of "indicators of sexual abuse." This allowed parents to find a simplistic explanation and a scapegoat for everyday problems that included: bedwetting, sleep disorders, etc. This in turn may have led parents toward inappropriate questioning of their children. Eichelbaum does not seriously discuss this more probable possibility in his report.

The fact that ACC provided lump-sum payments of $10,000 per allegation, for victims of alleged sexual abuse. Proof of abuse such as a conviction in a court of law was not required. This may have been a highly motivating factor for some parents to inappropriately question their children in order to ascertain if they might have the grounds to lodge a complaint. ACC paid more than $500,000 to around 40 parents of Civic childcare centre children, some receiving multiple payments of $10,000 (McLoughlin 1996).

Parent Meetings

The first formal parent meeting is often cited as having occurred on 2 December 1991. However, there was a significant earlier date. I recommend that anyone interested in all the details of these meetings read Hood, chapter six. There is a great deal more there, including much direct eyewitness testimony. Below is just some of the most relevant detail.

o    1st Meeting.

A Christchurch Civic Child Care Centre management meeting on 28 November 1991 (Hood pp 243 –246) included several key parents and Detective Eade. Ellis had been suspended just over a week earlier, it was two months before the first formal allegation of sexual abuse surfaced, and a further three months before the first formal allegation of sexual abuse that was to result in a conviction. The possibility of widespread abuse at the childcare centre was discussed. Reports regarding what some children had said were treated as allegations, although no allegations of abuse had yet been made. At the end of this meeting, it was decided to phone all parents about the concerns raised, without revealing details. It was consequently and unsurprisingly only a matter of days before every parent had a very exaggerated picture of what had been revealed. In fact, as Hood notes (p244), childcare centre parents were being asked to believe what adults were telling them, not children. At the close of this meeting, according to Hood (p244)

"each parent took away a list of names [of other Civic Centre parents]…over the next couple of days, they phoned them with an alarming and mysterious message: ‘there is going to be an urgent meeting…to discuss concerns that have come from the children’".

This fact alone indicates that parents were likely to have been comparing notes from very early on in the investigation.

o    2nd Meeting.

The first mass meeting was at the Christchurch Civic Child Care Centre on 2 December 1991 (Hood p249 –262). It was attended by parents who had laid complaints, city manager John Gray, the Police and some social workers including evidential interviewer Sue Sidey, who addressed the meeting. It began with an air of secrecy, although the parents had heard many rumours and the morning’s paper had supplied further detail. Ms Ruth made a dramatic entrance, bursting into tears immediately. Detective Eade told parents that there were "concerns," but refused to give details. This must have served to fuel the anxieties even further. Sue Sidey also spoke, telling parents to observe their children for indications of a whole host of common childhood behaviours. She clearly thought they were all indicators of sexual abuse, and the parents must have inferred this. She advised parents to read their children books like Katie’s Yucky Problem by Lynda Morgan and A Very Touching Book by Jan Hindman. Parents shouted, screamed and cried. Some left in disgust, saying that they did not want to be party to a witch-hunt. The effects of this meeting are indisputable and had huge bearing on the future course of the case; a core group of 13 parents who believed in the truth of the ‘allegations’ was immediately formed, and they began to meet regularly (Hood p255). The composition of this group closely matches that of complainant families. Yet this meeting was held prior to any credible allegation. In fact, by the end of that month, the Police investigation was closed down.

o    3rd Meeting.

This was at Knox hall on 31 March 1992. It may be described as infamous and for a description, I refer the reader to Lynley Hood (pp333 – 335). A handout given to parents that said, among other things: "Believe what they [your children] say." Parents were given Accident Compensation forms in order to lodge claims for $10,000, which at that time was the full and usual amount paid by ACC to anyone claiming to have been sexually abused.

This meeting had an effect on Ms X/Dogwood (Bander 1997). She had already heard that Ellis had been suspended as he was suspected of sexually abusing some children. Dogwood/Bander relates how she attended this meeting sure her son had not been abused. However, the nature of this meeting seems to have made her more ready to believe it may have happened.

"It was an unforgettable event…Why such a crowd? And why the police? There was a police car outside and several police were standing at the back of the hall…my impression of the meeting…was that it was a serious matter and there must have been some validity to the complaints…that’s when the seriousness occurred to me... the meeting made me distinctly uneasy.

Sally Ruth (Ms Magnolia / Parent D)

The first complaint turned out not to be an allegation at all. It came from a boy in November of 1991, who told his mother, he "didn't like Peter's black penis." Although he then explained that this remark was "only a story", his mother interpreted his statement as meaning that he had been sexually interfered with by Ellis.  His mother Sally Ruth, "parent D" (referred to as "Ms Magnolia" by Hood) had worked in the sexual abuse counselling area for most of her adult life and was a founder member of START, a private sexual abuse therapy organisation which emphasised "always believing the child" and working toward "unlocking all the pain, bad memories and feelings" of sexual abuse.  She also identified herself as a victim of childhood sexual abuse (Ansley, 1993).  This woman took it upon herself to contact other parents of children attending the Civic childcare centre, and co-ordinate sessions discussing possible abuse of their children by Ellis (Brett, 1993).

Ms Ruth appears to be a key source of contamination. She is referred to by Sas (p17) as the parent "informally assuming that role [of designated social worker]" for the parents. Ms Ruth was described by some parents as "a wonderful resource and support" (Sas p20). She

"... blatantly ignored the directive not to spread information…provided parents with literature…discussed other symptoms, phoned other parents...outlined all the places the children alleged they were taken…labelled the accused as a child hater and molester" (Sas p20).

There should be strong questions and doubt raised about the reliability and mental stability of Ruth (Magnolia/D). She was a woman who had "recovered "memories of abuse herself, had made a suicide attempt, and had in her own words a "politically motivated obsession" with collecting press clippings on sexual abuse cases. Ruth said (in 1992, Hood p403) that

"there’s a paedophile ring and they’re out to get us."

On page 225, Hood describes what clearly looks like an example of ACC fraud perpetrated by Ms Ruth.

Ms Ruth and the relevant concurrent Crèche investigation (1992)

The parent Ruth went on, in September 1992, to make similar allegations about another male crèche worker, at another crèche. This was after the closure of the Civic childcare centre but before the deposition hearings. Subsequent police investigation resulted in no charges being laid.

Hood (2001 pp403 - 406) gives a detailed account of this investigation. The investigation was handled very differently to that of the Civic childcare centre:

o    Police interviewed the named worker the day after a formal complaint by Ms Ruth.

o    There were more detailed staff records at the second crèche.

o    After one formal interview with the child, undertaken two days after the formal complaint, no allegations were made, and the accused went back to work.

o    The crèche supervisor directly discussed the rumours with parents on a one to one basis. "Nothing was hidden, but there was no public pronouncement," the supervisor told Hood.

o    When Ruth and her friends tried to arrange a meeting for concerned parents, Detective Sergeant Bob Hardie had the meeting cancelled. Hardie had worked on the Ellis case.

o    Hardie advised the crèche supervisor that she could continue to employ the named worker after Colin Eade had told the worker that he was going to arrest him. He continued at work.

o    When Ruth began repeating allegations by telephone, and showing photos of the named worker to children involved in the Civic childcare centre case, Hardie advised the crèche administrator to threaten Ruth with defamation proceedings.

This woman was instrumental in passing on information between parents of the Civic childcare centre and clearly had a leading role in the development of the case. Her behaviour at the second crèche, as described by Hood (p403 – 407) and others, should raise serious doubts about both her credibility and her sanity. This consideration is relevant and discussion of it needs to be out in the open. Eichelbaum makes no mention of this in his report, nor did it feature in reports of the trial (eg Hood 2001, Newbold 2000). Sim (2002) however dismisses Hood’s revelations regarding Ruth on the grounds that Hood’s picture was "derived from descriptions of her rather than first hand experience". This is insufficient reason to dismiss Hood’s extensive and meticulous research.

The parents directly contaminated the interviews themselves. According to the 1999 court of appeal (introductory point no 7),

"Generally before an interview commenced there would be a short discussion between the interviewer and the parents covering any disclosure the child had made and their responses to it, and any behaviour they had noted, with possible explanations for it, and the child's background and friends and contacts with other crèche children".

The evidence for widespread contamination of the children’s evidence by parents is very strong. Some parents appear to have been panicked, and others seem to have pre-judged Ellis. As time proceeded the potential for the construction of inaccurate memories was very strong. As these factors can confuse matters in cases of abuse, it is highly desirable, if not essential, that allegations are detected early and evidential interviews are carried out promptly. In this manner the possibility and effect of contamination is minimised. That did not occur in the Ellis case.

Back to section contents

(ii)        Contamination by Children

It is fascinating that Zelas looks for consistencies between the different children’s stories as proof of veracity. In doing so she often succeeded in proving unreliability, as shown elsewhere in this report. The records state that parents were organising playgroups and sleepovers in order to "help them remember" (Hood p341). Davies’ claim (p35 Davies report) that "Children are not fed tid bits of information from other children" seems a strange thing to say, as the parents admitted to Hood (p341-345 and elsewhere), that they did just that.

Back to section contents

(iii)       Contamination by Police

For the first crucial three months of the police inquiry, Eade was in sole charge. There is evidence of visits by Eade to parents and of him speaking directly to the children during the enquiry. (Hood p 310, 269) There can be little doubt that from the start, Eade passed on the strong message that Ellis was guilty of terrible crimes, and would go to jail. This, delivered by an authority figure, is a powerful message to a young child and it can influence recall by way of memory construction through, amongst other factors, the child's use of schema. Z/Kari Lacebark had been exposed to this notion, from some source, as is evident in her 4th interview

Kari:    Yep, now I’ll just have to get the beds ready and the jail ready and stuff. Should Peter have an empty jail?

Sidey:  What do you think?

Kari:    I think he should have a bit of furniture.

Sidey:  A bit of furniture, okay.

Kari:    But he was really bad. When will it be decided?

Sidey:  What?

Kari:    Um, for Peter to go to jail or not.

Sidey: Um, I don’t know. That’s up to [male name]. Okay now … um, where else

have you been with Peter?

The search by police for cages, special trapdoors, etc, also indicates that they took seriously allegations of a great ring of Paedophiles and satanic ritual abuse (SRA) operating in the Civic childcare centre. Additionally, the police neglected to interview Ellis and the other Civic childcare centre staff until very late in the investigation. The delayed interviews were more an interrogation and arrest, rather than a search for the truth. There is a clear bias operating in the input of the police (see also section (1.5.iii) A Priori Bias).

Back to section contents

(iv)       Contamination by Child Therapists

The NSW royal commission in 1997 reported that therapeutic intervention could contaminate a child’s evidence. This creates a problem if it is desired that (for whatever reason) therapy take place.

Loftus (1993) refers to a number of studies into contamination of evidence by therapists. For example, (p526) she talks of therapists who don’t take no for an answer, encouraging clients to ‘guess or tell a story’. She quotes research that shows that

"... the simple act of imagination makes an event subjectively more likely."

On page 527, Loftus quotes another study (by Mulhern 1991) that shows "the alarming discrepancies that often exist between therapists’ accounts of what they have done in therapy and what is revealed in video or audiotapes of the same sessions." In her discussion of how false memories can be constructed (p533) Loftus states that

"... false memories [are] created …with techniques that are not all that different from what some therapists regularly do – suggesting that the client was probably abused because of some vague symptoms, labelling a client’s ambiguous recollections as evidence of abuse, and encouraging mental exercises that involve fantasy merging with reality."

Some children involved in the Civic childcare centre case received therapy under the supervision of Karen Zelas before there were any convictions. In at least one case, therapy sessions preceded the completion of the child’s evidential interviews (Hood p343, Eichelbaum p9). There is a strong likelihood, in this case, that the therapists involved treated the children under the assumption that they had been abused. This may have strengthened any existing belief, on the part of the child, that they had been abused. It would seem to be exceedingly difficult if not impossible to treat a subject for sexual abuse without affirming to the subject an acceptance of the occurrence of the said abuse. As earlier noted, some of the children who gained convictions had parents who were therapists working in the sexual abuse field. Eichelbaum failed to properly consider these issues, there is hardly any discussion or information on the issues in the report

Back to section contents

(v)        Contamination by Adult Counsellors

The widespread dissemination of panic and bias is well illustrated by the story of one ex-Christchurch Civic Child Care Centre worker, Sandi (Hood’s pseudonym, described p 424). Sandi had not been accused, or arrested, but was pursued relentlessly by police "[the police] visited Sandi seven or eight times, and stayed for an hour or two each time.". When she sought counselling for her fears that the police might remove her child from her care, she was told that she was "having trouble coming to terms with the other women’s guilt." This is evidence of at least one therapist’s belief in the accused guilt, well before the trial. She also described how she found it difficult to find a counsellor who was available, as they were very busy with the Civic childcare centre families.

Back to section contents

(vi)       Contamination by social workers

Ms X/Dogwood (Bander 1997) reports that she had been assigned a social worker, who had visited Bart several times. The social worker also came to visit her one evening and talked to her for an hour after another "disclosure" by Bart. It seems possible that such social workers spread information between various agencies.

Back to section contents

(5.2)     The significance of delay between the alleged events and the allegations

Using information from the Eichelbaum report, the delays between the dates of the alleged offences that resulted in convictions and the time of the first formal interviews range between 7 to 50 months (see the table in my appendix 2). Recall becomes less accurate over time, especially for young children (Loftus 1979, etc). Ceci and Bruck, in their 1993 review of the scientific literature discuss Oates and Shrimpton’s 1991 study and state that it shows that

"... the effect of the delay of interview was especially consequential for the misleading action questions; children interviewed after a long delay were more susceptible to suggestion than those interviewed after a short delay."

The time it took for most of the allegations to emerge from the multiple interviews suggests the strong likelihood that the children may have made use of that time to divine what it was that they believed the adults wanted them to say (see Ceci & Bruck 1999). Pointedly, even Sas (p11) admits that there is only a brief "window of opportunity" when reliable evidence may be obtained. Any reasons or excuses for the delay are irrelevant to the potential effects of the delay.

Ms X/Dogwood makers it clear that her son (Bart Dogwood) did not make any spontaneous allegation for some time, if at all (Bander (1997). She states that in March 1992 when the third meeting of civic parents happened, she still did not think "Tommy" (Bart) had been abused.

"He would have told me if he had been" (p9)…"I also felt sure…I would have noticed [if] something was amiss." (p27). "During his more than two years of attendance he had never given us the slightest clue that he had experienced anything untoward there." (p25)

It was not until two months after that meeting that "Tommy said something to his older brother George, that clearly indicated something serious might have happened to him at the Creche [Civic childcare centre]".

His first evidential interview was a few days later.

Even in ideal terms the only useful interview evidence was probably the interviewing of the children by their parents, given the up to 8 months delay before the "expert" interviews commenced, and given the ages of the children. The forensic value of these parent interviews was clearly non-existent.

I conclude that given this delay alone, it was too late for any allegations to be reliable. Other evidence would have to be found. Surprisingly, this issue was swept aside in the Eichelbaum report’s conclusions.

Back to section contents

(5.3)     The significance of the ages of the children

In his conclusions, Davies states that "young children’s memories are inferior to adults", and they forget more rapidly (p34). This is well backed up in research by Loftus and others (eg Lindsay et al 1991,). Ceci and Bruck ( 1993) review five recent studies that all confirm this. Furthermore, they state (p409):

"... our review of studies conducted during the first 70 years of this century indicated that almost without exception, researchers believed children were more suggestible than adults."

Foley and Johnson (1985) found that six-year olds are easily confused between memories of real events and memories of imagined events. This seems particularly apposite to this case. Younger children are also more susceptible to suggestion (Bruck and Ceci 1999, Ceci et al 1997). This has been known for a long time, Ceci and Bruck (1993) quote research going back to 1948 (Burtt).

Back to section contents

(i)         Infantile amnesia and memory recall in young children

Neath (1998) summarises the research on infantile amnesia on pages 352 – 353. He defines it as

"... the finding that most people cannot recall events that involved them prior to the age of about 3 or 4 years."

Many of our seemingly real early memories are really only constructs after the event from external sources, -such as parents repeating stories and discussing the family photo album. It is important to have objective information available in order to validate the memories. It is also slightly more complicated than it first appears, as memories may have existed before being lost. Neath asks, (p353) "why is recollection so poor or even nonexistent after age three? [for events that occurred before the age of 3]" He can find no theory that adequately explains all data.

Simcock and Hayne (2002) in a well controlled experiment using 80 subjects (aged between 2 and 4 years) have shown that young children can recall earlier autobiographical events after a six month delay, when tested through re-enactment. This performance drops considerably for younger children, aged around 2 to three years at the time of the event, if a one-year delay occurs before the test of recall. On verbal tests the recall is especially poor. They conclude that children’s verbal reports of an event are limited by their poor verbal skills at the time of the event. The authors state,

"... language development did not render these perceptually based memories accessible to verbal recall."

In other words, they cannot use the words they have since learnt to describe their old memories. This was a strong effect. To quote again,

"There was not a single instance in which a child used a word or words to describe the event that had not been part of his or her productive vocabulary at the time of encoding."

If this study is replicated, then it will mean that in interviewing children about their pre-school memories, as in the Civic case, a reliance on verbal reports is fraught with problems. In fact, if Simcock and Hayne are correct, wherever children use language other than what was used at the time of the reported event, it is unlikely that the statements are representing true memories. Their general finding that verbal memories are poorer in pre-schoolers is not new. Nurcombe (1986) found that pre-schoolers remember as much as adults when the task does not emphasise verbal recall.

Another finding in this study was that as children grow older the process of improvement in recall, both verbal and non-verbal is a gradual, continuous process, not a sudden, dramatic event as implied by the word amnesia.

The overwhelmingly consistent finding is that younger children’s memories are significantly less accurate than that of older children. This is probably due to both the maturation of the brain and the gradual development of strategies for memorising and retrieval. Memories become increasingly more reliable (but never perfect) up to the teenage years or later when, according to Ruffman et al (2001) the frontal lobes fully mature. This part of the brain is used for source monitoring and episodic memory. Neath again on p 354 discusses how few 5 year olds display particularly good rehearsal strategies (in this case multiple item rehearsal) when memorising a list of words. In a simple word recall experiment, there is a steady increase in accuracy with increasing age of subjects (in a comparison of 4, 7, 10 year olds, and adults).

Back to section contents

(ii)        Spontaneous false allegations in young children

Bruck and Ceci (1999) have shown that pre-school children are disproportionately vulnerable to a variety of suggestive influences. In their 1996 book published by the American Psychological Association, the authors found that

"... through suggestive interviewing techniques and repeated questioning, children can be led to get wrong not only peripheral details, but the central gist of events they experienced, even events affecting their bodies that could have sexual implications."(APA press release 1996).

This means that some children when asked direct (but not necessarily leading) questions, will falsely assert that they were touched sexually. These allegations may be volunteered without prompting, and so appear spontaneous.

In the Ellis case, at least 127 children were formally interviewed (McLoughlin 1996), 118 of them subsequent to the December 1991 parent meeting (Sas p 10), yet only a handful were presented to the court. In court many charges were dismissed or thrown out. After the convictions, one child retracted of her own accord. The small number of complainant children could have simply represented that small percentage that will make such false allegations.

Consider Hood’s thesis that the least credible evidence was systematically removed in the prosecution’s preparations for the trial. It took some time and much questioning by parents and therapists before the allegations emerged, although some parents deny this to be the case. The children who gave evidence at the trial may therefore represent the most suggestible of the at least 127 children officially interviewed. That could be compounded, to the degree by which these children were simply exposed to more interrogation, although an investigation into this could control for this effect.

Perhaps the term "spontaneous " is misleading. Ceci and Bruck (1993) point out that children may tell untruths (inadvertently or deliberately lying) when they are motivated to do so (p426). They may lie to avoid punishment, they may have overheard discussions of an event and want to gain acceptance in the group, they may fear reprisal and wish an avoidance of embarrassment.

Back to section contents

(iii)       Children in the courtroom

The Ellis trial operated according to our amendments to the Evidence Act in 1989 and the then new evidence regulations regarding the videotaping of child complainants. It is my opinion this allowed for children’s evidence to be regarded as equal to that of adults, and disallowed the usual face-to-face court challenges and cross-examination. Hood discuses this issue on page 111,

"Over the more than two years between the introduction of the regulations, and the start of depositions in the Christchurch Civic Crèche case, judges weighed the requirement to ‘minimise stress on the complainant’ against the requirement to ‘ensure a fair trial for the accused.’ More often than not, they ruled in favour of the complainant."

My judgement is also that this was what happened in this case. I am also unconvinced that facing a jury in person is necessarily stressful for young children. In the experience of young children, television is commonly associated with fictional presentations and unreal depictions (eg cartoons). It seems an inappropriate medium to use in a courtroom in these cases as children have problems distinguishing fact from fiction. They were questioned, or at least cross-examined by a man on a television screen, rather than someone in person. Hood page 484

"When each child witness was called, an attendant took the child and his/her support person to the video room... no one except the crier was allowed to enter the room... the three wigged and gowned men who appeared on the screen..."

Even the appearance of the judicial garb may have lent an air of surrealism to the children’s perception of the moment.

Back to section contents

(5.4)     Suggestibility in young children

The evidential interviews with X/Bart Dogwood have entered the public domain, with their publication by Barry Colman. The Eichelbaum report also contains analysis of them. Many suggestive questions were asked during these five interviews. This does not seem to be in accord with the standards of the time. Bart often changes his answers to accommodate these suggestions. This supports my assertion that perhaps the more suggestible children were eventually chosen to appear in court. The crown prosecutor acknowledged that some of his evidence was "bizarre and difficult to comprehend."

Loftus (Loftus and Calvin 2001) describes that our ability to accept or resist false information is variable in the population. The screening of children (from at least 127 to just 11 who came to court –to the seven who obtained convictions- to six after the retraction) may represent a similar selection process to a stage hypnotist when choosing suggestible volunteers. Perhaps only the children most susceptible to the misinformation effect (Loftus, Hoffman & Hunter 1989, Poole and Lindsay 2001, etc) passed through this screening process.

There is some scientific literature on testing for suggestibility. Endres 1997 for example describes six formal tests for suggestibility (p53). He discusses The Bonn Test of Statement Suggestibility (BTSS) (p54), this test is useful for schoolchildren as it is administered verbally, and includes both misleading and factually correct questions. It has been found to be reliable by other researchers with both school age and pre-school children. Tests like this are in forensic use in Germany.

Endres (p48) quotes evidence that

"... indicates that habitually anxious and inhibited persons with weak cognitive abilities tend to be especially vulnerable to suggestive interviewing." They may also be more imaginative, "Stable individual factors related to vividness of mental imagery and reduced standards of reality monitoring thus apparently play a role in the implantation of false memories."

It may still be possible to make an assessment of the children’s suggestibility, using the existing available data on the Civic childcare centre children (for example, a blind analysis of the videotapes by an expert who could evaluate the children in terms of suggestibility. It may be that suggestible behaviour patterns could be discerned, such as always agreeing with the interviewer, not volunteering information spontaneously, etc) If so, the results would certainly have some bearing on the case. In relation to changes in best practice, whilst it may not be routinely carried out here, in Germany at least, an analysis is often attempted of the intrinsic reliability of each child (i.e. assessing their suggestibility). I would recommend that we look at adopting the same practice here in cases where there exists a selection process of witnesses, no spontaneous allegations and a delay in conducting formal interviews.

A new test of children’s suggestibility has been announced by the American psychological Associatiuon (Scullin, Matthew H. Kanaya, Tomoe and Ceci, Stephen J. 2002). The authors used the Video Suggestibility Scale for Children to predict the performances of 50 pre-school children in terms of accuracy and their responses to suggestion. In their conclusions the authors state that the test

"may provide important information to forensic interviewers and psychologists about a child’s tendency to provide inaccurate information about true and suggested events…observing children to its [the test’s] questions can provide investigators with a sense of how sensitive some children are to suggestive influences.".

Back to section contents

(5.5)     Do the children’s "symptoms" of abuse have any scientific validity?

A number of professional bodies have come to the conclusion that there exists no cluster of symptoms that indicate sexual abuse.

o    "There is no single set of symptoms which automatically indicates that a person was a victim of childhood abuse. There have been media reports of therapists who state that people (particularly women) with a particular set of problems or symptoms must have been victims of childhood sexual abuse. There is no scientific evidence that supports this conclusion."

American Psychological Association, Questions and Answers about Memories of Childhood Abuse, 1995.


o    "Psychologists recognize that there is no constellation of symptoms which is diagnostic of child sexual abuse."

Canadian Psychological Association, Position Statement on Adult Recovered Memories of Childhood Sexual Abuse, 1996.


o    "Previous sexual abuse in the absence of memories of these events cannot be diagnosed through a checklist of symptoms."

Royal College of Psychiatrists, Reported Recovered Memories of Child Sexual Abuse, 1997. (UK)

Loftus (and others) discuss how adults seek psychotherapy initially because of symptoms. These symptoms are then interpreted as evidence of abuse even though there is no real evidence to suppose this, and a number of alternative explanations are just as likely. The problem is that people in general and therapists "have a tendency to search for evidence that confirms their hunches rather than search for evidence that disconfirms it." (ibid. p530) I contend that just these processes were operating in the Ellis case.

The prosecution claimed that various "symptoms" that the children exhibited were interpreted as evidence of abuse in the Ellis case (Hood page 521). The proponents of this theory included the parent Sally Ruth, the therapists who conducted their work concurrent with the investigation, and most certainly Zelas. This "evidence" would undoubtedly increase expectation (on the part of therapists, interviewers and parents) for "disclosures." From all accounts, the children displayed few signs of distress whilst attending the childcare centre. There are numerous contemporary reports of both children and parents liking Ellis. However, some time after Ellis had been dismissed from the Civic childcare centre, signs of disturbance, anger and unhappiness began to appear. A behavioural analysis would direct that one must examine what subsequent experience these children had in common that could be a cause. The obvious one is the experience of the evidential interviews. The "therapy" sessions could also have much the same effect, as they seem to have been conducted in a similar manner. A behavioural scientist would hypothesise that those children, who withdrew from the interview/therapy process, would show an overall improvement in these behaviours. The experts in the Eichelbaum report did not carry out that important analysis or even consider this possibility.

A list of behaviours said to be indicators of abuse was given to parents at one of their many meetings (both formal and informal) during the investigation. Sas (p16) states that this was inadvisable. I can find no scientific study that establishes a cluster of symptoms (or "syndrome") that reliably indicate sexual abuse.

It is interesting to note that Ms X/Dogwood seems to think Bart’s vomiting on one occasion was a sign of abuse (Bander p39). But she also reports that on the same evening, her baby had a tummy bug, and had also vomited.

Sas appends to her analysis of each interview a section entitled "Clinical indices of reliability" In essence, within these sections Sas is applying precisely the theory of such "symptomatology" to the Ellis case. Eichelbaum does not note this. Sas (p9) suggests that "nightmares, toileting problems, sexualised behaviours, unexplained fears and misunderstood comments" are likely to indicate to parents that sexual abuse has occurred. These behaviours are commonplace amongst normal, non-abused young children, and all have more likely non-abuse aetiologies.

Alarmingly, Zelas propagated this unsubstantiated theory in court. In one such occurrence (the Geoff Scott case, a male hospital crèche worker, aged 35, convicted in 1994) reported by the Dominion newspaper 3 May 1996, Zelas stated during the trial in Wellington;

"...excessive and persistent sex play by children was one sign that they had been subjected to sexual abuse."

In the same report, it is stated that,

"...she [Zelas] agreed that in most cases, a formal evaluation of the family situation and its effect, or otherwise, on the child had not been made."

This is fascinating in the light of Zelas’s admission to me a year earlier in 1994 that "there is no set of behavioural or emotional symptoms which reliably indicate abuse has occurred." (Harper 1994)

Another unpalatable fact, often difficult for some people to accept, is that children who have genuinely been sexually abused, may suffer no or little psychological trauma as a result, and are in fact free of any symptoms of trauma. Kendall-Tackett, Williams and Finkelhor (1993) in a review of 45 studies found that

" one symptom characterised a majority of sexually abused children… One third of victims had no symptoms."

Increased trauma seems to occur when others spell out the depravity of the crime to the child, frequently counsellors and/or interviewers (Loftus 1993 p534). Ergo, a healthier approach might be to down play the evilness of the deed to the child, and emphasise that it may not be difficult to get on with one’s life. Exactly the opposite seems to have occurred with the Civic childcare centre children.

McNally's book ( 1998 pp. 22-26) describes the violent condemnation by the US Congress in 1999 of an article by Rind, Tromovitch and Bauserman in a prestigious psychology journal which came to the same conclusion based on a meta-analysis of 59 studies. Apparently this kind of information is not always palatable to politicians and is not considered in "treatment" or "therapy" programmes.

Back to section contents

(5.6)     Is consistency proof of reliability?

The prosecution claimed (Hood p523) that the children had given consistent central detail in their allegations. This was presented as evidence of reliability. A British Home Office paper quoted by Eichelbaum (p44) prepared by psychologists Davies and Westcott in 1999 concludes

"No reliable method yet exists of judging the truth or otherwise of children’s allegations, either from their demeanour at interview, or from features of their accounts."

It appears to me that the jury, due to Zelas’s testimony, may have believed the assertion that since the children’s stories were consistent in regard to "central issues", then abuse was more likely to have occurred. Zelas’s use of the term "consistent with" is highly ambiguous and one cannot really be sure exactly what she means by this. Nevertheless, the jury may have placed weight on these statements without fully understanding them. The jury was not well directed in that they were not made aware of the fact that children can produce allegations that are convincing but false.

Not withstanding, I contend that the central allegations made by the children were not consistent. The prosecution eliminated references (in the charges) to Ellis’s Hereford Street address at the trial. That was surely to avoid alerting the jury to inconsistencies in the children’s evidence as to this central fact. Panckhurst (Hood p542) in his appeal pointed out that

"The final seven children named 21 other children as observers or participants in their abuse, but none of the named children confirmed the allegations."

My contention was supported by the New Zealand Law Journal editor (2002 p1) who stated,

"No one who has read the confusion and contradiction displayed by the witness statements that Hood recites can be happy that the convictions are safe."

When the Crown prosecuting lawyer Stanaway referred to inconsistencies in central detail, (Hood p479), he said this was because the child was talking about different episodes of abuse. This is nevertheless an admission of unreliability. It would weaken the effect of an adult’s evidence, and there is no reason for it not to equally weaken that of a child. Eichelbaum concludes (p119) that there was a "body of broadly similar allegations." Sas had noted (p58) that "it was difficult to know at times if children were talking about different occasions and incidents." The perhaps more likely interpretation of the inconsistencies might be that they indicated unreliable evidence, evidence that may have had no base in reality. This was not adequately explored in the Eichelbaum report. Stanaway contradicted himself when he claimed (Hood p479) that young children remember central detail better than peripheral detail (the data does not support this statement).

The jury may have given Stanaway’s explanation for the confusion between different incidents undue weight. Williamson erred in not noticing or stopping another possible source of bias in manoeuvres made by the prosecution. For a number of the charges Prosecutor Stanaway manipulated, between the depositions and the trial, the central details of who committed the crimes, and where they were committed (Hood pp 479 - 481). This was probably done to avoid alerting the jury to the original thrust of the Civic childcare centre charges, that the women were co-offenders. The women had been discharged at this stage. This was the prosecution’s plan B. It appears to be a dishonest move because it hid inconsistencies in central details of the allegations. Eichelbaum fails to discuss this in his concluding remarks. In fact, he seems to state that as the Court of Appeal had already considered the case twice, no further analysis of such matters was appropriate (Eichelbaum report p120).

Back to section contents

(5.7)     Coercion

The prosecution claimed in the case of Child X/Bart Dogwood (Hood p522) that Ellis had terrorised the children into silence. Thus, the continual questioning and encouragement was necessary until X/Bart felt safe to tell. The evidence that Ellis told the children ‘not to tell’ is contradictory and weak. (e.g. on page 16 of Davies’ report it is claimed that child Q/Lara Palm believed for some time that Peter would turn her into a gherkin if she told, and there exist reports that Ellis was liked by the children at the time when he was at the childcare centre). McLoughlin (1996) states,

"By most accounts, Ellis was extremely popular with children and parents alike."

McLoughlin claims that, "experts agree that such threats make disclosure of abuse more likely". Ceci (quoted by McLoughlin) discounts the theory that Ellis (if guilty) would have succeeded in keeping them quiet about the acts that he was found guilty of committing. He states,

"It, in my experience, is exceedingly unlikely that you can coerce a group of children this age into silence for prolonged periods of time when the following were allegedly involved… anal penetration, forcing children to walk over precarious ladders perched high above buildings, defecating and urinating on children… these are events which cause almost instant revulsion in children, night tremors, unwillingness to go to school, fear of the perpetrator.

In my view it is very very unlikely that you could persuade children to be silent about that for long periods and also to assert affection for the perpetrator which many of these children did. So on that level I’m exceedingly sceptical. I don’t think the bizarre stuff happened. Does that mean nothing happened? Well, I simply don’t know. No one else knows either except God and Mr Peter Ellis."

If Ceci is correct then the prosecution’s theory is weak indeed, and did not justify the long period of probing. Furthermore, just how Ellis had threatened the children to keep them silent, and had managed to keep the threat alive by some form of remote control has never been explained.

McLoughlin says he sent Ceci’s 14 pages of comments to Attorney-General Paul East, who passed them to the Crown Law Office, which apparently gave it to Social Welfare. Neither organisation has publicly challenged Ceci’s findings.

The onus of proof that Ellis effectively kept the children silent for months or years surely must lie with the prosecution. They do not appear to have done this.

Back to section contents

(5.8)     A statistical analysis of the reliability of the experts’ evaluation of the evidence

Koehler (2001) discusses the lack of understanding of Bayes’ statistical law among health professionals. The reader might like to test his/her knowledge of it.

Assume 1% of all women who undergo a breast screening test have breast cancer;

The screening test has an 80% success rate of correctly identifying those who have cancer that it tests. If a patient has cancer he/she has an 80% chance of this being identified.

The test has a 10% false positive rate (i.e. 10% of the patients who are free of cancer will be incorrectly identified as having it). This means that of those without cancer 90 % will be correctly cleared of the disease.

But for those who have been diagnosed as having cancer, what is the likelihood, or probability, that they do in fact have it? The answer appears below.

Somewhat surprisingly, at first glance, the probability is 0.074. In other words there is a 7.4 % chance of the said patient having cancer and by simple logical implication there is a 92.6 % likelihood of the same patient not having cancer. Yet, according to Koehler many professionals would highly over inflate the likelihood of cancer being present after one doctor’s positive diagnosis from this test, and others like it (for example a test for HIV infection.)

Here is the reasoning toward the result:

Say 1000 patients are tested. As 1% of those tested have cancer this means 10 have the disease. The test correctly identifies 80% of the cancers it screens. This means 0.8 x 10 or eight of the 10 patients with cancer are identified…So far the doctor is doing well…. 

However 990 out of the 1,000 do not have cancer and a 10% error rate of false positives results in 100 of these patients being incorrectly identified as having it. So we have a total of 108 positive diagnoses, only 8 of which are correct. The percentage probability of a positive diagnosis being right (correct positive) is 8/108 x 100 = 7.4 %

So, given these conditions, if the doctor tells you that you have cancer, there is a 92.6 % chance that in fact, you do not have cancer. One should not however lose sight of the fact that the patient now has a seven-and-a-half fold higher chance of having cancer than has a non-screened person in the general population.

Second opinions, and other tests are then required. The same sort of analysis could be applied to prosecution witness Zelas’s apparent assessment that the children’s testimony was reliable. The only device Zelas seems to have used to determine reliability of the allegations themselves was an examination of "consistency of central detail." This device does not give the impression of being as reliable as a breast-screening machine. Yet, she certainly never stated that she thought that there was a low probability that the allegations were unreliable. The second opinions in this case are divided. Sas agrees. Those who disagree include the defence expert Le Page, psychologists Dr Ceci and Barry Parsonson (who viewed many of the videotapes). Davies said he would not pronounce on the issue, presumably because he had insufficient information to make a reliable assessment.

Some early studies showed that clinicians couldn’t determine whether or not a child has been sexually abused. (eg Dawes 1989). This was verified by Ceci et al (1997), who found that professionals were "unable to distinguish true from false accounts" from children’s testimony through interviewing children who possessed some true, and some inaccurate (implanted through suggestive techniques) memories. These studies provide an indication of the rates of true and false positives.

In other words, in the Ellis case, the rate of false positives (false allegations being officially accepted as true) must be close to an absolutely appalling 50%, as seen by simply from viewing the videotapes. In order to begin to approach the legal requirement of beyond reasonable doubt, other factors must be considered, such as gullibility of the witnesses, (not considered in the Ellis case), delay between alleged events and formal allegations, and the ages of the children. But in the Ellis case, as I discuss elsewhere in this document, these factors are negatively correlated with accuracy in this case. For example, accuracy of recall is poorer for pre-school children compared with older children. Considering this, with Zelas’s apparent confidence in court and Eichelbaum and Sas’certainty, one is forced to conclude that this trio is most likely deluded in regard to their ability to be sure the children’s evidence was reliable. This effect also has consequences whenever the terms of reference of the Appeals and Inquiry is considered, as whenever the terms are narrowed this effect is probably inflated.

Ziskin and Faust (1997) have examined the reliability of psychiatrists and psychologists in courtrooms. They found many are ill fitted to this work because they are used to being sympathetic and helpful to their clients or patients. This is illustrated in the Christchurch case of Simone Doublett, who said that she lied about satanic ritual abuse. She stated (in 1995) that her therapist, psychologist Dr Lynne Haye was not good at telling that she was lying, and that she lied to gain attention. Haye now accepts she was gullible, She had attended a seminar, and read literature that had convinced her SRA was real and not uncommon. According to Ziskin and Faust the accuracy of psychiatric judgement is poor, and did not improve with experience, age or special qualifications. In the courtroom, the need to uncover truth is a task that may conflict with this role. They also concluded that human behaviour has so far resisted a solid theoretical foundation that can make reliable and accurate predictions.

Zelas falsely denied in court having given a television interview with Paul Holmes about the case (Hood P503). I find it unlikely that she could forget participating in a major television interview. Whatever the reason, this is not the action of a reliable witness (court transcript p318 Hood).

For four reasons, Eichelbaum did not exercise sufficient caution in agreeing with Sas’ overconfident assertions regarding the reliability of the children’s evidence. He ignored this evidence of Zelas’s unreliability, he over estimated Sas’ reliability, he did not consider the statistical significance of the other second opinions that differed, and his ignorance of the use of statistics led him to overestimate the reliability of experts in general.

This suggests a need to train judges is some basic aspects of diagnostic statistics and relevant findings from the science of psychology. Koehler discusses the implications this type of overconfidence in expert diagnosis, and gives an example of a counsellor who had dealt with a patient who had received a positive result from a test for AIDS. This led to the suicide of a patient who did not have AIDS, but was led to believe that he/she had the disease due to the one positive test result. An educated use of statistical formulae by the health worker would have alerted the patient to the unlikelihood of having the disease despite the single positive test result.

Back to section contents

(5.9)     The effect on reliability of the proportion of impossible and unlikely allegations

There were many clearly impossible allegations in this case. For example, Ellis did not own a giraffe, there were no cages hung from ceilings, he did not turn a frog into a cat, no children visited the Masonic lodge and he didn’t dig up Jesus. Other allegations that at first appeared believable were later found to be highly unlikely (Hood p 435). This includes all the events alleged to have happened at Ellis’s Hereford Street house. The Crown apparently convinced the jury that Ellis had cleverly blended tricks and fantasy into the abuse so that the children would recall the two together (Hood p521, 522). A fair analysis of this case would consider all the allegations arising from all of the at least 127 children officially interviewed. The proportion of the allegations that were unlikely, impossible or probably untrue should have been calculated. This information would have been helpful to the jury in assessing the credibility of the evidence as a whole. This is a standard scientific practice. It is improper to only report positive results because the total set of data could simply be a random set of numbers signifying nothing. Unfortunately, this seems to be a standard legal (advocacy ) practice. As already discussed, this is precisely how many unsupported theories of psychic abilities arose. The jury also needed data from control groups, that is, studies where children were questioned in a similar fashion, but it is known beforehand that no abuse took place. As I have also already shown, there exist control data in regard to the level of false allegations from children in such situations. Such a comparative analysis with the Civic children has not been done to my knowledge.

Christchurch Press court reporter Van Beynan (1995), who attended the trial, discusses complainant Z/Kari Lacebark whom he calls Lucy. In regard to this child, four out of four charges laid resulted in convictions (see appendix 2 for a comparison with the other children who appeared as witnesses). Van Beynan states,

"almost every allegation she made was undermined by detail she provided…the activities she described...were not physically possible…Throughout her six interviews Lucy changed her story about what offences occurred and their frequency. She gave differing accounts of the state of her and Ellis’s clothing [etc]…As confusion reigned, anatomically correct dolls were introduced to the interviews which produced further conflicting accounts. In the first interview Lucy asked 14 times for the questioning to stop …everything divulged by Lucy was open to the alternative hypothesis that she was in essence making it up as she went along…"

If Lucy was at the high end of the jury’s reliability scale, it raises the question of where does this leave the testimony of the other children upon which convictions were based?

Back to section contents

(5.10)   The retracted evidence

N/Zelda Cypress (referred to as S by McLoughlin) dramatically retracted her allegations 18 months after the trial. She said she had lied about the abuse, because "she thought everyone would be mad [angry]" had she told the truth. (Hood p549). Hood notes that our courts, and prosecution expert witnesses encourage us to believe the child when making allegations of sexual abuse, but when a retraction occurs, we are told not to believe the child.

The following points are important in any consideration of the retraction.

o    In the trial the mother admitted reading to her three books about sexual abuse from the age of three.

o    Zelda "was obviously very fond of him" (Ellis) according to the mother.

o    It was three years after last attending Civic childcare centre before the mother (Ms Cypress) made her initial allegation. The child was nine years of age at her first formal interview.

o    In common with most interviews, anatomically correct dolls and repeated and leading questions were used.

o    The initial disclosure occurred when Zelda’s mother asked suggestive questions and described Ellis in a negative light.

o    Her first disclosure matched the scenario of one of her mother's initial questions.

o    Throughout the interviews, when making allegations, she hid her face, and tried to slide under the table. Ellis relates to Hood that he recognised this as typical of her when she knew she was not telling the truth.

o    During the first session (of six in total), Zelda asked 14 times that the interview stop.

o    When her statements did not fit the details of the accepted scenario, the interviewers ignored them.

o    Many of the activities that she described were not physically possible.

o    She gave conflicting accounts of the state of her clothing, the state of Ellis' clothing, the location, and the perpetrators during her interviews.

o    She alleged that Ellis drove a car to transport children to another location.

Eichelbaum accepted that this girl could be "in denial", meaning her allegations could still be true. Given the preceding points, this seems most unlikely.

Others retracted before the trial (Hood p435). Detective Eade himself stated (Hood p585) on TV in 1997 "I’d be surprised if not all of them have done it [retracted] at some stage," This comes as a revelation after reading the conclusions made in Eichelbaum report. No analysis is made in the Eichelbaum report of any comparison of the interviews of N/Zelda Cypress with those of the other children who obtained convictions. Such a comparative study would be a useful means of determining the reliability of much of the evidence. It should be remembered that prior to retraction, Zelda’s evidence was regarded by the jury as being the most credible of all of the complainants (see the appendix 2). If Zelda’s retraction does not differ in any significant way from the other retracted complaints then, assuming the retractions are genuine, doubt is cast upon all the evidence.

Back to section contents

(5.11)   Repressed Memory and Child Sexual Abuse Accommodation Syndrome

The whole case could be seen as a special case of the repressed memory theory driving the investigation. This overwhelmingly discredited theory could lead to bias and unreliable evidence with children in the same way that it operates with adult abuse "survivors." The effect of the delay on reliability is similar (the delay may have been shorter here) as in a classic case of a "recovered memory". The techniques used to attempt to "recover" memories appear similar to those used by the interviewers in this case. To quote Loftus (1993) on repressed (later termed "recovered" memory);

"The therapist convinces the patient with no memories that abuse is likely, and the patient obligingly uses reconstructive strategies to generate memories that would support that conviction. These techniques can be found in numerous autobiographical accounts."

Robbins (1998 p5) also likens a SRA scare to a case of repressed memory.

However, Karen Zelas did not consider this to be the case. Along with Sas and the prosecution she appeared to believe that the children could remember all the while, but were reluctant to "disclose" due to embarrassment and/or threats made by Ellis. These factors and many others are explained by the CPM in the "Child Sexual Abuse Accommodation Syndrome" (CSAAS). This syndrome was first defined in a thesis by sexual abuse worker Dr Roland Summit (1983). Summit himself admits that his theory is based on a perception rather than hard data. According to Goodyear-Smith (1993) CSAAS has been summarised in the following way

o    Sexually abused children tend to contradict themselves

o    Sexually abused children cover up the incident

o    Sexually abused children often show no emotion after the event

o    Sexually abused children often wait a long time before making their accusations.

Although it is entirely possible that sexually abused children may exhibit some or all of the preceding it is scientifically invalid to use such behaviours as diagnostic of abuse as all the behaviours can occur in non-abused children.

The forgoing tenets all appear and are applied to the various children in the report of Sas at some point. Acceptance of the syndrome on Zelas’s and Sas’s part can explain much of their interpretations of the children's responses and acceptance of bad interviewing techniques.

Back to section contents

(5.12)   Comparison with overseas cases

The recent Shieldfield nursery case in Newcastle, (England) strikingly bears much resemblance to the Christchurch Civic Child Care Centre case. There can be little doubt that no actual abuse occurred. The two accused workers were acquitted and later awarded damages in a libel case. Justice Eady awarded the two 200,000 pounds each, which was "the highest permitted level." He said:

"What matters primarily is that they are entitled to be vindicated and recognised as innocent citizens who should, in my judgement, be free to exist for what remains of their lives untouched by the stigma of child abuse."

The Wee Care Nursery case (or Kelly Michaels) in America (quoted in Ceci and Bruck 1993) is another. The 115 convictions against Kelly Michaels involving 20 children were eventually quashed.

The similarities of the Wee Care case with the Christchurch Civic Child Care Centre case include:

o    An unsubstantiated allegation from a parent that sparked off a great deal of concern among other parents.

o    Initially, the publication of unsubstantiated rumours in the media. In his official judgement, (paragraphs 146 to 321), Justice Eady lists endless biased media coverage of the Newcastle case. He cites headlines such as, "branded abusers" and "report reveals scandal of child abuse."

o    The interviewing of a large number of children, as a result of a belief that there were a large number of victims. Even after Ellis’s trial, the police continued to believe 80 children had been abused at the Civic childcare centre (Ansley, 1993). The number of children was dramatically reduced when it came to the trials, suggesting a selection of more suggestible children, (or, but less likely in my opinion, a selection of the children who really had been abused).

o    The clear existence of belief in SRA among some parents, and others involved in the case. In the Wee Care case, this involved the supervisor playing a musical instrument in the nude, and making children drink urine, and eat faeces.

o    No alleged act was witnessed by any adult or reported by the children to their parents at the time.

o    No parent noticed the smell of urine or faeces on their child or signs of genital soreness at the time.

o    The suppression of obvious SRA allegations in the charges that made it to trial, so as to unfairly render the remaining evidence more believable.

o    The presence of bizarre and unlikely allegations in the children’s testimony.

o    Parents were asked by social workers to watch for signs that could indicate abuse had taken place. Eg: bedwetting, nightmares, masturbation, any changes in behaviour, etc.

o    Questionable child evidential interviewing techniques.

In the Wee Care case, Ceci and Bruck (1993) state (p405),

"... there was repeated provision of an atmosphere of accusation with interviewers informing children, - it’s OK to tell…you’ll feel better once you tell…most of the children were told by the interviewers prior to their own disclosures that their peers had already disclosed that Kelly Michaels was a bad person who had hurt them."

Back to section contents

(5.13)   Conclusions

Charges were tried in regard to eleven children and their ages ranged between 4 – 9 years when they were first formally interviewed. Their ages had to be less than five years when the offences were said to occur or they would have been at school. My analysis, from the data in my appendix 2, puts most of the children at less than four and a half years of age at the time of the alleged offences.

It is perfectly clear to me that the memories of the young children in this case were unreliable regardless of how they were interviewed. This is particularly relevant because of the predominantly verbal nature of their reports. The one attempt to get around this problem using anatomically correct dolls has proved to be an unreliable alternative. Ideally the children should have been interviewed at the scene of the alleged crimes, and not in a new environment, as contextual cues have been proven to have a significant positive effect on recall. (Neath 1998 pp 119 – 126). Furthermore, the time delays clearly are cause for extreme concern about their reliability, given their ages. Significantly, the one child (among those who gained convictions) who was old enough to be reasonably reliable, retracted her evidence later. This girl (N/Zelda Cypress) was nine at the time of being interviewed, although she was asked to recall events said to have occurred over four years previously.

During the delay before the interviews, there is ample evidence of widespread contamination of the children’s memories by parents, other children, therapists, adult counsellors the interviewers and police. In respect of Detective Eade the New Zealand Law Journal (Feb 2002 editorial) is most scathing,

"... the investigation ... suffered from a clear fault which was that it was driven by a junior officer with a bee in his bonnet."

It is still being asserted (by Sim 2002) that the Law Commission’s report (preliminary paper no26 quoted in Sim ibid. p4) is correct in stating,

"... there is currently no evidence to support the proposition that children spontaneously and without prompting fabricate claims of sexual abuse…"

One could debate the meaning of the term fabricate; however, the tenor of this statement appears at odds with the studies quoted here. Perhaps the science is just not getting through to our judicial system.

Eichelbaum failed to seriously consider the limited power of children in regard to understanding the difference between fantasy and reality. He also failed to acknowledge, in his reasoning, information supplied by Davies as to the ability of children aged under 7 years old to reliably discern issues regarding temporal accuracy.

Eichelbaum’s apparent failure to properly consider memory research is of very serious concern. It brings into question the whole procedure of appointing judges to review matters where knowledge of psychology is of crucial importance.

I believe his poor selection of experts, neither of who was an experienced research scientist in the relevant areas, contributed to this failure. These relevant areas of expertise should have included the expert having conducted childhood memory research.

So the indications are that the conditions surrounding the formal interviews made the convictions very doubtful.

To summarise:

o    The children’s evidence was thoroughly contaminated by their parents, other children, therapists, and the police.

o    The six children examined in the Eichelbaum report were less reliable than adults or older children. They were possibly a selected group of the six most suggestible of the over 127 children originally interviewed

o    These children were unlikely to have any accurate and reliable memories of anything that happened at the Civic childcare centre by the time they were formally interviewed.

o    No credible evidence of coercion to maintain secrecy about alleged events emerged

o    The experts who claimed that the evidence was reliable had questionable expertise.

o    There are no clusters of symptoms that reliably indicate sexual abuse.

o    A broader examination of evidence than that presented at trial shows little consistency of central detail in the children’s accusations.

o    Acceptance of the theory of denial must have a bearing upon the reasons that the retracted evidence and quashed convictions were (illogically) adjudged to have little or no bearing on the remaining evidence and convictions.

o    Similarities with overseas cases was not examined adequately

Eichelbaum failed utterly to thoroughly consider the compelling evidence for the forgoing conclusions. The New Zealand Law Journal editor (p1, February 2002) nicely sums it up.

"... either Sir Thomas did not read the statements [of the child witnesses] because, like everyone else he restricted himself to the filleted evidence that the judge allowed in, or, with respect, his judgement is at fault." …

Ceci and Bruck conclude their 1993 review of the scientific literature (for the respectable Psychological Bulletin) by stating,

"It seems particularly important to examine the conditions prevalent at the time of a child’s original report…If the disclosure was not made in a non threatening, non suggestible atmosphere, if the disclosure was not made after repeated interviews, if the adults…are not motivated to distort the child’s testimony through relentless and potent suggestions and outright coaching, and if the child’s original report remains consistent over time, then the young child would be judged to be capable of providing much that is forensically relevant. The absence of any one of these conditions …ought to raise cautions in the mind of the court."

As will be shown in the following section, those cautions were not raised adequately, nor heeded in the Ellis case. They should have been made, according to this evidence, most forcefully indeed.

End of Section Five

Next Section
Previous section
Back to section contents
Back to document contents

site stats