Testing the Reliability of Child Witnesses

Stephen Ceci wants an answer to what seems like a pretty simple question: How do you know when a child is lying?

Ceci is a professor of developmental psychology at Cornell University and a nationally recognized authority on the reliability of child testimony. It's an interest that has taken him to the core of some of the most heart-wrenching court cases in U.S. history involving unspeakable allegations of child abuse.

The popular proverb that children are basically honest and always tell the truth went out the window following a series of sensational child abuse cases in the 1980s and 1990s that hinged primarily on the testimony of children.

Some adults who were clearly innocent watched their lives come apart at the seams when accused of heinous crimes against children, even if they were later exonerated. And, no doubt, some scoundrels escaped punishment because a child's testimony could not be believed.

Testing ‘Suggestibility’

Ceci and a couple of colleagues are developing a test that they believe could help investigators determine if a particular child is especially vulnerable to being swayed by questions asked by a social worker or investigator. They hope to come up with a scale showing how "suggestible" any young witness really is.

"There's lots of research that says younger children are more suggestible than older children, but that isn't terribly helpful because they want to know about a particular child who's making allegations," Ceci says. "Is this child more suggestible?"

That's a question that few raised in the 1970s when the federal government passed a well-intentioned law aimed at clamping down on child abuse.

The law gave states a financial incentive to force teachers and medical personnel to report any evidence of child abuse. That led ultimately to what the Los Angeles Times called the "lynch-mob syndrome."

No one set out, of course, to create a monster, but that's what eventually happened. It all came to a head in the 1980s with sensational charges against a child-care center in the Los Angeles suburb of Manhattan beach.

Seven adults at the center ultimately faced more than 200 charges in the McMartin preschool case. A series of trials, which ended in either acquittals or hung juries, dragged on for six years, costing taxpayers more than $16 million. Numerous lives were destroyed along the way, and it all finally collapsed because too many of the children simply could not be believed.

Some claimed they had witnessed a beheading, and were forced to drink blood, and subjected to horrible sexual abuse. Some even swore they had seen one of the adults at the center fly through the air.

‘Gun-Shy’ Prosecutors

A subsequent Grand Jury report laid the blame for the failed prosecution on the prosecutors themselves. Interrogators had been so aggressive, even to the point of telling some children they were "dumb" if they hadn't seen the things reported by some other kids, that in the end many of the children couldn't separate fact from fiction.

Other cases, including one a short time later in San Diego, had similar results. A child was snatched from her family, and her father forced into counseling for sexual abuse that he clearly did not commit. He was exonerated two years later when DNA evidence proved that a man who was in jail for raping another child was the perpetrator.

These sensational cases led to a reluctance to prosecute child abuse charges because of fear of failure, according to the National Center for Prosecution of Child Abuse, an affiliate of the National District Attorneys Assn. Prosecutors, according to one expert, became "gun-shy."

Much progress has been made in the last decade, Ceci says, but the situation still cries out for some means of determining which children are most likely to be swayed by the questions of an interrogator.

So Ceci and Matthew Scullin, now an assistant professor of psychology at West Virginia University, and Cornell doctoral candidate Tomoe Kanaya developed a test to see if they could come up with some measurement of a child's "suggestibility."

Video Trial

They produced a benign 10-minute video, showing children at a birthday party, and showed the video to 50 children with a median age of 4.5 years.

"There's some things that happen in the video, like a toy gets broken, a fire alarm goes off. They have cake and ice cream and one kid kind of messes up and drops cake on her lap," Ceci says. "Things like that."

Following the video, the children were asked 18 questions, some of which contained false or suggestive information, such as "when Andrew broke the toy, was it an accident or did he do it on purpose?"

"But it was not Andrew who broke the toy," Ceci says. "It was a different child."

The children were questioned again, this time a bit more aggressively by an interviewer who told them they had made some mistakes.

The children were rated on how often they gave the correct answer, and the number of times they changed their answers after being told they had made mistakes. Most of them did quite well, making only a couple of minor mistakes, but some of the children clearly were more suggestible than others.

Subsequent tests, over a four-week period, confirmed that those who had been identified as most suggestible wandered farther from the truth as the interrogation intensified.

"So it was pretty predictive," Ceci says. The test appeared to pinpoint which kids needed the most care during the interview.

But how would it work in the real world?

Interviewer Is Key

Most children called to testify have been traumatized. Would a simple test about a birthday party really tell which children are most likely to lie?

The researchers don't know yet, because that part of the project is the most difficult to carry out. Obviously, they can't traumatize kids to see how they react, so Ceci is trying to set up a program at a major hospital where the test can be administered to kids who are having a rough time, like suffering from an infection in a private area.

That will require parental approval, of course, and it's still less traumatic than a case of child abuse, but it's about as close as the researchers can get to the real thing without risking additional harm to the child.

Ceci says he's "fairly optimistic" the test will work even for children who are going through a difficult time, but in the end whether or not the test helps will depend largely on the integrity of the interviewer.

If the interviewer is a vigilante, the kid doesn't have a chance.

"If you keep coming at very young children, a very, very high percentage of them will break down and tell you what they think you want to hear," Ceci says.

Most children can be swayed, even to think the unthinkable, when at the mercy of an adult who's out to get someone. And that is a terrible predicament for prosecutors, parents, and of course, children

Lee Dye’s column appears weekly on ABCNEWS.com. A former science writer for the Los Angeles Times, he now lives in Juneau, Alaska.