Accuracy and reliability are key to the best lie detectors.
Accuracy Defined
It’s common to speak of accuracy, which is the ability of a system to classify truthful and deceptive persons correctly. It is often expressed as a percentage of correct classifications. The terms accuracy and validity are used synonymously in scientific literature.
Specificity is the True Negative rate or percentage of truthful persons a system correctly identifies as credible. Sensitivity is the True Positive rate or percentage of deceptive persons a system correctly identifies as not credible.
Reliability
Reliability is a prerequisite for accuracy; it refers to the consistency and stability of a procedure when it is repeatedly applied under the same conditions.
EyeDetect has demonstrated near 100% reliability in independent studies because it fully automates the testing and scoring process, eliminating the subjectivity and variability inherent in manual systems. Essentially, it analyzes test data and scores tests consistently, every time.
Manual Testing
Manually administered and scored credibility assessments are inherently problematic. They rely on human examiners to administer tests and interpret physiological data.
This introduces potential for human error, subjectivity, examiner fatigue, and bias, resulting in inconsistent variations in testing and scoring based on individual skill and preference. (Swift, Spake & Kohia, 2016)
Objective Reliability
EyeDetect and VerifEye are fully computerized systems that monitor involuntary eye movements and other physiological responses, with sophisticated algorithms performing the scoring. This automation ensures near-perfect reliability across three critical dimensions:
- Standardized Protocol: The examinee experience is identical for every individual, as the system administers all testing procedures.
- Elimination of Bias: Algorithmic data analysis and scoring remove all potential for human fatigue, bias, or interpretation errors, ensuring results are based solely on data.
- Data Integrity: Test data is encrypted and tamper-proof, preventing administrators from altering results and ensuring security from capture to final report.
Certified Advantage
EyeDetect and VerifEye hold a significant scientific advantage by incorporating base rates of deception (prior probabilities) into their scoring algorithms. This factor—the expected frequency of deception within a given population—allows the system to calculate a statistically more accurate posterior probability of deception. (Nelson, Handler & Thiel, 2021; Warner et al, 2024).
Manual scoring typically ignores this crucial condition, which introduces a known statistical weakness and a significant potential for error by treating every examinee as having a neutral (50/50) chance of deception, regardless of the context.
The shift from a subjective, variable-dependent manual assessment to a computerized, algorithm-driven system provides a level of standardization and objective reliability that manually scored tests cannot match, solidifying its position as a superior method in forensic science.
References
- Andersson, M., Boateng, B. G., & Abos, O. J. (2014). Validity and Reliability: The extent to which your research findings are accurate and consistent. Research Gate.
- Nelson, Handler, & Thiel, C. (2021). Posterior Odds of Deception and Truth Telling for Low and High Prior Probabilities. Research Gate.
- Swift, Spake & Kohia. (2016). Examiner Fatigue and Ability to Concentrate in Objective Structured Clinical Examinations for Physical Therapist Students. Journal of Allied Health, 45(1), 62-70.
- Warner, Anderson & Kircher. (2024). Bayesian Decision-Making Mitigates Effects of Base Rates on Outcome Confidence: A Monte Carlo Simulation. Psychophysiological detection of deception test & Forensic Credibility Assessment, 53(1).
- Webb, Hacker, Osher, Cook, Woltz, Kristjansson & Kircher. (2009). Eye Movements and Pupil Size Reveal Deception in Computer Administered Questionnaires. Foundations of Augmented Cognition. Neuroergonomics and Operational Neuroscience.
Converus (“with truth”) provides scientifically validated credibility assessment technologies that help protect countries, corporations and communities from corruption, crime and threats.