ANN ARBOR—A new University of Michigan study reinforces general reservations about the reliability of using lie-detector tests in cases of alleged sexual abuse of children.

“Professionals involved in cases of alleged sexual abuse are desperate for certainty,” says U-M researcher Kathleen C. Faller, professor of social work. “Error in either substantiating or not substantiating sexual abuse can have devastating consequences.

“A false positive can lead to grave injustice not only for the accused, but also for the child. On the other hand, a false negative may doom the child to continued sexual abuse and retaliation for having dared to disclose, and may destroy any remaining hope the child has for protection.”

In examining 42 child sexual abuse cases in which the accused perpetrator underwent a polygraph test during the course of investigation, Faller found that lie-detector results are unrelated to other signs of sexual abuse, such as the child’s statements or demonstrations of abuse, medical evidence, psychological symptoms or indicators of sexual abuse from sources other than the child.

In other words, additional evidence supporting sexual abuse existed in these cases, regardless of whether an alleged offender passed or failed a polygraph test.

According to the study, when alleged offenders pass a polygraph, criminal prosecution is not sought. However, failing a lie-detector test does not necessarily lead to prosecution, as 10 persons in the study who failed police polygraphs were not prosecuted.

“Upon reflection, this should be an anticipated outcome,” Faller says. “Polygraph findings are not admissible in court. Demonstrative evidence of sexual abuse has to be provided.

“The best predictor of criminal prosecution was corroborative information from sources other than the child, the aggregate of hard evidence, such as medical findings, witnesses or other victims, and police evidence.”

Faller found no consistent pattern in what child protective services rely upon in deciding whether to verify allegations of sexual abuse. Health and mental health professionals seem to rely on factors other than lie-detector results in their decisions to substantiate such claims.

Finally, Faller cautions that her findings have limitations because the majority of cases in her study come from one site, the sample is of modest size, and many of the cases were referred to the study because polygraph results and findings on other indices of sexual abuse were discrepant in the first place.

More information: 

• Kathleen C. Faller
• bernied@umich.edu