Research asks: Just how reliable are children's memories

By Tammy Hardiman

Until recently, it was believed that a child's memory of an event was sufficiently unreliable to allow it to be used as testimony in a court of law. It is now becoming clear that children are sometimes the only witnesses to events, and it's important for their testimony to be used ­ not only in police investigations, but in the courtroom as well.

In fact, children are increasingly being allowed to testify about events without corroboration from an adult witness. As a result, there are numerous people in the justice system who want to know to what extent they can depend on what a child is saying; in other words, how valid is the testimony? Dr. Carole Peterson, a professor of psychology at Memorial, is interested in determining the extent to which knowledge gained about a child's memory in a controlled laboratory setting can be applied to a real-life situation. For the past three years, Dr. Peterson -- with a grant from the Natural Sciences and Research Council (NSERC) -- has been conducting research to determine whether children remember events accurately when they are traumatized or distressed.

Children recruited

In carrying out this research, several of Dr. Peterson's students spent time at the emergency room of the Dr. Charles A. Janeway Child Health Centre in St. John's recruiting children who were willing to participate in the study. To be eligible for the study, which began in 1993, the children had to have gone to the Janeway's emergency room with injuries classified as trauma, which are mostly lacerations requiring stitches, and broken bones.

More than 200 children participated. They were interviewed at their homes approximately a week after the injury, then at six-month and one-year intervals. They were asked questions about what happened during their injury and their hospital experience. "As there were no laboratory controls we didn't know exactly what happened to each child," Dr. Peterson explained. "The only information available was from adult witnesses of the child's injuries, and for many of the children there were multiple witnesses. The adults always said the same thing, and there was never a disagreement between what the adults said and the hospital records. Thus, the adult reports were used as our 'gold' standard against which we compared the child's information for accuracy." As well, the children's parents -- who were usually the adult witnesses -- were asked to rate how upset their child was on a scale of one to six, with one meaning "not upset at all" and six meaning "extremely upset." Among the children in the study group, there was a wide range of distress experienced at the time of injury.

Pre-schoolers' memories

The age group that Dr. Peterson was most interested in was pre-schoolers, ages two to five, since the credibility of pre-schoolers' memories is most often questioned.

"The results after the second interview, six months after the injury, had some surprising results," Dr. Peterson said. "We were surprised at how much the children remembered. Two-year-olds only remembered about half of the important information, but by three years of age children remembered most of the event. The thing that was very surprising was how incredibly accurate their information was. Only about 10 per cent of what three-year-olds said was inaccurate. By this age, the children virtually never made mistakes in terms of the important things that happened to cause their injuries."

Similar inaccuracies

Dr. Peterson observed that the children almost always gave inaccurate information about the same things: the identity of the person who came to their rescue, and what happened after they left the hospital. "When they were hurt -- blood was going everywhere and they were screaming -- the person who got to them first was what they made mistakes about," she said. The children's accuracy was almost unaffected by emotionality; how upset the children were and their level of distress made virtually no difference to what they remembered. Children who were absolutely hysterical were just as accurate as those who were less upset. During the interview process, it was discovered that the methods used to question pre-schoolers is important.

"Young children tend to say 'no' to yes/no questions regardless of the question, because they don't understand what this type of question means. They don't yet possess the mental faculty to evaluate the truthfulness of the question and then verify this truth. For these children, yes/no doesn't mean true or false; rather, they use it in terms of 'I don't feel like talking' or 'I want toplay,'" said Dr. Peterson. "Unfortunately, many people think these are the easiest questions to answer, thus they are the ones that children are generally asked."

Specific questions best

Children respond well to questions which are more specific and the answers they give are very accurate. The "wh-questions" -- who, what, when, where, and how -- are more easily understood and answered by children. "A child will give absolutely accurate information with a wh-question, but the same question asked as a yes/no question will provide inaccurate information," said Dr. Peterson.

Pre-verbal memories

Another facet of the study involved children of about one to two years of age who could not verbalize what happened to them. Dr. Peterson was interested in whether they could talk about the experience when they did acquire language six to 12 months later. Generally children who experience a traumatic event while still pre-verbal can't access the memory once they begin to speak -- this phenomenon is known as infantile amnesia. Some children in the study seemed to show non-verbal evidence that they remembered the incident, but later there was little recollection of the event expressed either through words or behavior.

One year after the initial injury the children were re-interviewed and many of them were asked misleading, suggestive questions -- the kind of questions often asked in court -- to see whether they would assimilate this information as part of their memories. Many of the children will be interviewed again two years after their injury. The results of these interviews are currently being gathered; nevertheless, the results to date suggest that children accurately remember forensically important information. They remember what caused their injury and are usually only inaccurate about factual information such as who reached them first once the injury occurred. These results could have important implications for testimony in courtrooms of the future.