The reaction time three-item Concealed Information Test :: effects of information leakage, deception and anxiety
Loading...
Files
Date
2016
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
University of New Brunswick
Abstract
The reaction time three-item Concealed Information Test (RT-CIT) is a type of recognition test that examiners can use to detect concealed awareness of crime details among crime suspects. Examiners assume that only perpetrators and investigators know crime-related details and infer examinees are “guilty” if the test indicates that they possess crime-related knowledge. However, it is possible for innocent examinees to learn crime-related details from other sources (e.g., media, investigators), a phenomenon known as information leakage, which was the focus of the current study. Results revealed that the RT-CIT successfully differentiated between examinees guilty of a mock crime, innocent but informed of the crime (informed-innocent), and innocent but unaware of crime details (uninformed-innocent). Classification accuracy using a classification procedure (the Compound Classification Procedure, CCP) was also good for guilty (92.5%) and uninformed-innocent groups (100%). However, consistent with predictions, elevated false positive rates were found among informed-innocent groups (50%), supporting the conclusion that the RT-CIT is vulnerable to information leakage. Another variable of interest was the type of instructions given to participants prior to the RT-CIT, which prior research has found influences RT-CIT performance. Examinees in the current study read one of two instruction sets (deception, or control instructions) prior to completing a RT-CIT, but did not show any differences in RT-CIT performance. Individual differences in anxiety were also examined in an attempt to predict the accuracy of the RT-CIT. Preliminary analyses pointed to state cognitive anxiety as the best predictor, which ultimately accounted for significant variance in the accuracy detection score (the difference in response accuracy between irrelevant and crime details). Given strong performance of the RT-CIT, state cognitive anxiety did not account for significant variance in the RT detection score, RT and accuracy sensitivity indexes, nor did it add to the prediction of guilt beyond the results of the CCP. Taken together, the results support using the RT-CIT to detect concealed knowledge in criminal investigations, given that examiners take precautions against information leakage. However, given the artificial nature of mock crimes, the small predictive value of cognitive anxiety, and the questionable effectiveness of the instruction manipulation, more research is necessary.