Accuracy demonstrations, threat, and the detection of deception: Cardiovascular, electrodermal, and pupillary measures

This 1981 laboratory study examined how subjects' beliefs about polygraph effectiveness and threat of punishment influence detection accuracy on the Control Question Test and Concealed Information Test using a mock crime paradigm. Results demonstrated that perceived polygraph effectiveness significantly enhanced detection rates on the CQT but had weaker effects on the CIT, with skin resistance emerging as the most reliable physiological measure.

A mock crime experiment where participants either committed or did not commit a staged theft, followed by polygraph examination using both CQT and GKT formats with accuracy demonstrations manipulating subjects' beliefs about polygraph effectiveness (0, 1, 2, or 3 successful detections shown) and threat of punishment as experimental conditions. Physiological measures included skin resistance, heart rate, and pupil diameter.

This landmark study investigated psychological factors influencing polygraph detection beyond physiological measurement, specifically examining whether an examinee's perceived beliefs about test accuracy affect physiological responses and detection outcomes. Using a factorial design with mock crime scenarios, participants underwent both CQT and GKT examinations after being shown varying levels of demonstrated polygraph effectiveness and facing potential punishment threats. The Control Question Test demonstrated clear dose-response improvement in detection accuracy as subjects believed the polygraph more effective, while the Concealed Information Test produced less consistent results. Skin resistance proved the most diagnostic measure overall, though pupil diameter showed 80% accuracy for innocent participants on the GKT but only 33% for guilty participants, with different patterns on the CQT. These findings suggest that subjects' expectations and threat perception substantially influence physiological responses independent of actual deception.

The study demonstrates that polygraph detection accuracy is partially determined by examinee beliefs about test effectiveness rather than physiological responses alone, suggesting that pretest information and contextual factors significantly influence outcomes in operational settings. Findings highlight differential effectiveness of CQT versus CIT formats under psychological manipulation and support continued investigation of pupillometry and skin resistance as diagnostic channels, with important implications for examiner training, pretest procedures, and understanding the mechanisms underlying detection accuracy.

Background & Context

The 1981 Bradley and Janisse study emerged during a crucial period in polygraph research when scientists were examining the psychological factors that influence detection accuracy beyond mere physiological measurement. While much early polygraph research focused on identifying which physiological channels were most diagnostic, less attention had been paid to how an examinee's beliefs about the test's accuracy might affect their own responses.

This research addressed a fundamental question that continues to resonate in polygraph practice: Does the perceived effectiveness of the polygraph itself influence detection outcomes? The study also investigated whether external threats or punishment would affect physiological responses, and importantly, explored pupillometry—the measurement of pupil diameter changes—as a relatively novel polygraph channel alongside traditional measures.

The work built upon Lykken's influential Guilty Knowledge Test (later renamed the Concealed Information Test) and compared it directly with the Control Question Test, examining whether these psychological manipulations would affect both techniques similarly.

Research Design & Methodology

The experiment employed a mock crime paradigm where half of the participants committed a mock crime (stealing a dollar and concealing it in their pocket) while the other half remained innocent. Following the crime phase, all subjects were examined using both Control Question Test (CQT) and Guilty Knowledge Test (GKT) formats, with skin resistance, heart rate, and pupil size recorded as physiological measures.

The key experimental manipulation involved accuracy demonstrations conducted prior to the actual polygraph examination. Before the polygraph test, the effectiveness of the physiological detection apparatus was demonstrated to each subject, and they were led to believe they had been detected in either none, one, two, or three of three trials during this demonstration. This created four experimental conditions varying in subjects' beliefs about the polygraph's effectiveness.

A second manipulation involved threat of punishment, where some participants were threatened with negative consequences for failing the test. The study thus employed a factorial design examining:

• Level of demonstrated polygraph effectiveness (0, 1, 2, or 3 out of 3 detections)
• Presence or absence of punishment threat
• Guilty vs. innocent status
• Two questioning techniques (CQT and GKT)

Results & Key Findings

The accuracy demonstration manipulation significantly affected detection outcomes: subsequent detectability of subjects varied as a result of the manipulation with the Control Question test such that detectability increased as the level of demonstrated effectiveness increased. This represented a dose-response relationship—the more effective subjects believed the polygraph to be, the more detectable they became.

The key findings revealed important differences between the two testing formats:

• Control Question Test: Detection accuracy improved systematically as subjects were shown higher levels of polygraph effectiveness during the demonstration phase
• Guilty Knowledge Test: Results were less clear, with the accuracy demonstration manipulation having weaker effects
• Skin resistance was the most efficacious measure in both tests, while pupil and heart measures varied in detection accuracy depending on the test employed
• For the concealed information test, pupil diameter reliably discriminated between guilty and innocent participants with 80% accuracy for innocents but only 33% for guilty participants; for the CQT, pupil diameter did not reliably discriminate between the groups

The threat of punishment manipulation produced unexpected results. Threat of punishment did not affect detection results but did alter heart rate change such that those threatened, whether guilty or innocent, received rank scores more in the guilty direction than those not threatened. This finding suggested that external threat affects cardiovascular arousal non-specifically, potentially reducing the diagnostic value of heart rate measures.

Discussion & Significance

The Bradley and Janisse findings illuminate a critical psychological mechanism underlying CQT effectiveness: the polygraph's perceived credibility directly influences its actual accuracy. This supports the theoretical rationale that the CQT operates partly through psychological intimidation—guilty examinees who believe the test works will experience greater fear when lying to relevant questions, enhancing physiological differentiation. This mechanism appears specific to the CQT format, which relies on fear of detection.

The weaker effects observed with the Guilty Knowledge Test make theoretical sense. The GKT (now CIT) is designed to detect recognition of crime details through an orienting response rather than fear-based arousal. Because recognition is a more automatic cognitive process, it should be less susceptible to manipulation through believability demonstrations. This finding actually validates the GKT's theoretical foundation as a recognition-based rather than emotion-based technique.

The differential effectiveness of physiological channels is also significant. Skin resistance outperformed pupil diameter and heart rate, consistent with the broader polygraph literature emphasizing electrodermal activity as the most robust channel. The poor performance of pupillometry in the CQT format (though better in GKT) suggested that pupil responses may be more sensitive to cognitive processing demands than to arousal from lying per se.

Limitations & Considerations

The study employed a laboratory mock crime scenario with university students, raising questions about generalizability to real-world polygraph settings where stakes are genuinely high. The mock crime itself—stealing a dollar—involved relatively low emotional investment compared to actual criminal investigations. Additionally, the accuracy demonstration manipulation, while clever experimentally, doesn't directly parallel field practice where examinees' beliefs about polygraph accuracy are shaped by cultural factors and pre-test interviews rather than controlled demonstrations.

The threat manipulation did not produce the expected effects on detection accuracy, only on heart rate patterns. This negative finding limits conclusions about how external stressors interact with deception detection, though it does provide valuable information about the non-specific effects of anxiety on cardiovascular measures.

Practical Applications

This research provides empirical support for the importance of the pre-test interview phase in polygraph examinations. Examiners who effectively establish the credibility and accuracy of the polygraph during pre-test may enhance detection accuracy—not through deception, but by creating appropriate psychological conditions for guilty examinees to exhibit stronger differential responses. This finding validates established polygraph protocols that include detailed explanations and demonstrations of the instrument's capabilities.

However, the findings also raise ethical considerations. If polygraph accuracy depends partly on examinees believing it works, this creates a feedback loop where perceived effectiveness enhances actual effectiveness—but only for certain test formats. The weaker effects on the GKT/CIT suggest that recognition-based tests may offer more stable accuracy independent of psychological manipulation, potentially making them more suitable for contexts where examiner credibility establishment is limited or ethically constrained.