To reach their conclusions, researchers from the University of Colorado at Colorado Springs and Cornell University in New York recruited 41 children aged 3 to 7 years old and had a magician visit their classes. Later, the children were interviewed about their memories of the event by professional police interviewers.
In particular, the researchers were looking for any effect the kind of question asked--either a biasing question or a non-biasing question--had on the children's responses.
The results were surprising. "We discovered that children were more likely to disagree with a biasing question than disagree with a non-biasing question," said Livia Gilstrap, Ph.D., assistant professor of psychology at the University of Colorado. "That was the exact opposite of what we expected."
However, researchers found one important exception. If the biasing question was inaccurate (e.g., "Was his name Michael?"--when his real name was John), and the inaccurate information was being said for the first time, children were more likely to agree with the biasing question than to disagree with it.
Researchers also found that adults were more likely to ask biasing questions when the child was already disagreeing with earlier questions. In other words, if a child said "No, his name wasn't Michael" to the previous example, the interviewer was more likely to press the child with biasing questions.
Regardless of the questions being asked, however, researchers found that children who had just agreed with a question were likely to continue agreeing, and children who had just disagreed with a question were likely to continue disagreeing.
"Taken together, our findings suggest it might be the children, not the interviewers, who are actually driving the interview," notes Dr. Gilstrap. "If an investigator is asking a child questions about alleged crimes and the child is disagreeing with the investigators assertions, these data suggest that the investigator will tend to ask biasing questions of the child." Additionally, she noted, the data suggests that if investigators ask inaccurate biasing questions for the first time, children tend to agree with the inaccurate statements. "Although these data are still exploratory," Dr. Gilstrap notes, "this raises the concern that children who are reluctant to talk will be asked questions that are more likely to distort their reports."
Summarized from Child Development, Vol. 6, Issue 1, Reconceptualizing Children's Suggestibility: Bidirectional and Temporal Properties, by Gilstrap LL (University of Colorado at Colorado Springs) and Ceci SJ (Cornell University). Copyright 2005 The Society for Research in Child Development, Inc. All rights reserved.
Journal
Child Development