The purpose of this detection of deception experiment was to study the assumption of the Guilty Knowledge Test that subjects with guilty knowledge will be classed as guilty by the test regardless of their actual guilt or innocence. Prior to a polygraph examination, three groups of innocent subjects were given the same crime-relevant information as members of a group guilty of a mock crime. These innocent subjects either witnessed the crime, were told the crime details, or carried out innocent activities involving crime-relevant information. An additional group of innocent subjects had no crime-relevant information. Analysis of the Guilty Knowledge Test results showed that the detection scores of guilty subjects were higher than those in any of the innocent groups. In fact, with the exception of the innocent activities group, the innocent informed subjects did not differ from those in the uninformed group. The major conclusion is that subjects may have crime-relevant information and not be classed, based on the detection scores, as guilty.