||The current study evaluated the Computerized Screening System (CSS) for ports of entry. Additional primary objectives included evaluating the Relevant Comparison Test (RCT) for use at ports of entry, less invasive alternatives to skin conductance and the cardiograph, and alternative statistical methods for classification. Data were collected in two phases. Complete sets of recordings were obtained from 169 Phase 1 participants and 185 Phase 2 participants (N = 354). Participants were either guilty (n = 230) or innocent (n = 124) of committing a mock crime. Guilty participants transported a substance that appeared to be illegal drugs (n = 119), or they transported a device that appeared to be a bomb (n = 111). When the participant reported to the laboratory, a research assistant initiated a computer program that presented prerecorded auditory instructions and test questions to the participant. The computer administered a test entitled the Relevant Comparison Test (RCT) that was developed specifically for this project. The RCT included 12 relevant questions about the bomb condition, 12 relevant questions about the drug condition, and 24 neutral questions. Respiration, electrodermal, cardiovascular, and pupil reactions were recorded continuously throughout the test. As expected, guilty participants who transported the drugs reacted more strongly to questions about the drugs than to questions about the bomb. Participants who were guilty of transporting the bomb reacted more strongly to questions about the bomb. Innocent participants reacted similarly to questions about the drugs and the bomb, although there was a tendency for some innocent participants to react to questions about the drugs. Increases in diastolic blood pressure and systolic blood pressure were most diagnostic of group membership behind skin conductance. Contrary to expectations, pupil measures did not perform as well as skin conductance measures, and traditional discriminant analysis was more effective than the computer-intensive bagging and boosting classification techniques.