In a guilty knowledge test (GKT) paradigm, subjects were induced to attempt to deceive the experimenter on two separate occasions while electrodermal activity (EDA) was measured. For 20 of the 30 subjects, overt behaviors (changes in eye movement and facial expression) were recorded during the second GKT. Results indicate that the incidence of behaviors decreased during the subjects' deceptive responses. This behavioral inhibition coincided with increases in skin conductance level. These results support DePaul et al.'s hypotheses that the inability to find consistent nonverbal behavioral correlates of deception may reflect that observers are looking for the occurrence of deception-relevant behaviors as opposed to the omission of behaviors. Also, behaviors or the lack thereof may only surface during the brief period of skin conductance increases. Although EDA is a far more reliable predictor of deception than changes in eye movements and facial expressions, the behavioral data can provide an additional information source in seeking to detect deception in a controlled setting. 14 references. (Author abstract modified)