Some children lie when reporting sexual abuse. Most don't.

A trained investigator who has no prior relationship with the child can usually tell the difference during the interview stage, said David Raskin, professor of psychology at the University of Utah.When sexual abuse of a child is suspected, a therapist should turn the investigative interview over to a trained interviewer, said Raskin.

"It's important to get the witness' (child's) statement, not the interviewer's," he said. "You must obtain an interview that can be analyzed. This is not therapy, and therapy should always be done by someone else.

"You must find out as early as possible what is the truth."

Speaking during Grand Rounds at the Western Institute of Neuropsychiatry Tuesday, Raskin said a properly conducted investigative interview can substantiate an allegation or screen out fictitious reports.

Raskin is internationally known for his forensics expertise, particularly with the polygraph.

Sometimes, he said, physical evidence substantiates an abuse claim. But "clear physical evidence is usually lacking. The child is usually the only witness. We use special techniques to ensure accuracy."

In a forensic interview, the memory, language and cognitive level of a child are carefully considered. The goal is to "elicit a complete and accurate description of the events." Leading questions are only asked to test whether a child is readily suggestible.

Raskin said the fact-finding interview should be taped, preferably video-taped.

The United States lags behind Europe in using the technique, although Raskin said he and others have done research for almost 10 years and conduct workshops to train others.

The interview follows a sequence. First the child is asked for a free narrative of what happened, followed by open questioning, like "Tell me more about it." Direct questions can clarify information. The interviewer then asks questions to check suggestibility. The discussion always ends on a positive note. "You always want the child to leave feeling pleasant," Raskin said.

"It usually takes 20-30 minutes. It's amazing how much information a child can give you. It doesn't take weeks and weeks of play therapy, anatomically correct dolls" etc. "That's not only the slowest route, it's the most contaminated route."

Analysis of the interview includes consideration of whether the account sounded like a genuine memory or something of "invention or fantasy," whether the child had a motive or was pressured to make up the story, and whether the child "could or would be able to fabricate a statement with these particular qualities."

Younger children provide fewer details. The investigator analyzes the child's account with a "validity checklist."

A field study has been "encouraging," Raskin said, indicating the method may be reliable more than 80 percent of the time.

"We need to be able to gather evidence in a systematic way to protect children and society. It's important to identify the small minority of cases which are fictitious and need to be screened out."