This paper examines whether a non-intrusive computerized system that analyzes handwriting can detect deception in health care. Health systems are required to deal with false information given by some patients about their health (malingering). Studies have shown that clinical ability to detect deception is limited, and evidence suggests that better results can be achieved by using assessment tools than by relying on human detection. Currently, tools for detecting deception are intrusive and therefore less suitable for the clinician–patient relationship. Within-subject experimental design compared deceptive writing with truthful writing of 98 participants aged 21–36, recruited from the University of Haifa. They wrote true and false sentences about their medical condition on a paper affixed to digitizer that was part of a computerized system. Deceptive and truthful writings for all the subjects were compared. In the next phase, using profile analysis, subjects were divided into three groups according to their handwriting profiles, and the differences between deceptive writing and truthful writing of each profile were analyzed. Deceptive writing was found to be broader and took longer to write than truthful writing. Three distinct profiles were emerged, and significant differences in specific spatial and temporal measures were found for each profile. Preliminary results provide a unique perspective on detecting deception with this computerized tool. Possible applications for the health system and other fields, such as human sorting and internal security, are discussed.

Published in Cognitive Computation, 2014, 6(4), 849-855.