Research findings by the head of the university's psychology department, Professor Harlene Hayne, suggest there is a "strong risk" their evidence was contaminated by the way the interviews were carried out.
Addressing the Innocence Project conference in Wellington at the weekend, Prof Hayne urged the courts to consider the case again.
Ellis, convicted in 1993 of 13 charges of sexually abusing children at the Christchurch Civic Childcare Centre, has always protested his innocence and is fighting to clear his name. He served two-thirds of his 10-year prison sentence.
After a High Court trial and two failed Appeal Court hearings, a ministerial inquiry conducted by former Chief Justice Sir Thomas Eichelbaum in 2000 found that the interviewing of children who gave evidence at Ellis' trial was appropriate and hadn't been undermined by contamination by others.
However, Prof Hayne's research shows that questions put to the Christchurch children were much worse than those done in a similar American daycare child abuse case in the 1980s that Sir Thomas accepted in his report was a major miscarriage of justice.
Prof Hayne analysed hundreds of pages of verbatim transcripts of the pre-trial interviews with very young children involved in the Christchurch case and in a similar case involving New Jersey daycare worker Kelly Michaels.
In his 2000 ruling, Sir Thomas declared the Christchurch interviews were "best practice" and much better than those in the Michaels case, which he accepted were badly flawed.
Kelly Michaels, aged 23 when she was arrested in 1985, was a daycare worker at the Wee Care Nursery School in Maplewood, New Jersey. After a trial that lasted almost a year, she was convicted of 115 counts of child sexual abuse and jailed for 47 years.
She was released on appeal after five years, with the New Jersey Supreme Court ruling that the "interviews of the children were highly improper and utilised coercive and unduly suggestive methods".
Prof Hayne found the Christchurch children were each subjected to an average 400 questions by Social Welfare Department specialist service staff, compared with an average 200 questions in the Michaels case.
"There is a strong risk that the children's evidence was contaminated by the way those interviews were conducted," Prof Hayne told the Innocence Project conference in Wellington.
"The courts should look at this again."
Each child in the Ellis case was asked an average of 20 suggestive questions, Prof Hayne said.
"Just one can give a wrong answer in laboratory tests."
She said the accuracy of children was "terrible" when they faced large numbers of questions.
The 400 questions each child in the Ellis case was asked was double the number faced by children in the Michaels case, yet Sir Thomas had said the Christchurch questioning was best practice, Prof Hayne noted.
"The standard of the questions was not consistent with the best practice of the time, as a lot was known then about the dangers of suggestive questions," she told the conference.
"The standard of the questions in Ellis was not substantially better than those in Michaels."
Prof Hayne said as a scientist it was difficult to talk about certainty.
"We will never know what really happened," she said. "We can't reinterview (the Christchurch children) now as they will have taken on well-ingrained false memories of things that didn't happen.
"This case has been a tragedy for the children."
The Innocence Project, a group of scientists, writers and lawyers who seek to investigate possible cases of wrongful conviction, is a joint venture between Victoria University and Otago University.
