A comparative analysis of polygraph with other screening and diagnostic tools

Abstract : The purpose of this study was to conduct a limited review of literature published between January 1986 and May 2001 concerning the accuracy and reliability of screening and diagnostic tests in polygraph, medicine, and psychology. Out of 5,189 hits produced by the literature search, 1,158 articles and abstracts were reviewed, 145 were found to useful resulting in data on 198 studies. For field screening assessments, the sensitivity of polygraph, medical, and psychological tools was .59, .79, and .74 respectively. Specificity of polygraph, medical, and psychological tools was .92, .83, and .72. Agreement was measured with kappa. Among readers in polygraph, medicine, and psychology kappa was .77, .56, and .79 respectively. Reports in the literature of polygraph's accuracy and reliability (agreement) on specific issues appear to be consistent with published studies on medical and psychological assessment tools. However, there is an enormous range of accuracy and agreement not only within polygraph but also medicine and psychology. Although there were very few polygraph screening studies, accuracy reports were lower than those in medicine and psychology.

Systematic literature review of 1,158 articles yielding data from 198 studies across polygraph, medicine, and psychology, using sensitivity, specificity, and kappa coefficient as standardized psychometric measures to enable cross-disciplinary comparison of diagnostic and screening tool performance.

Background & Context

This 2001 study commissioned by the Department of Defense Polygraph Institute emerged during a critical period when polygraph testing faced increasing scrutiny regarding its scientific foundations. Medical and mental health professionals routinely use assessment procedures with comparable or even inferior psychometric properties without controversy, yet polygraph remained uniquely subject to criticism.

The study addressed a fundamental question in forensic science: how does polygraph accuracy compare to established diagnostic tools in medicine and psychology? By systematically reviewing 15 years of literature across three distinct fields, Crewson sought to contextualize polygraph performance against other widely accepted professional assessment methods. The polygraph approach involves a human reader using technology to measure and interpret physiological conditions and responses to make a diagnosis, very similar to the mechanics involved in diagnostic radiology.

This comparative framework was particularly important for policymakers and practitioners who needed empirical benchmarks to evaluate whether polygraph testing met reasonable standards for professional diagnostic tools, rather than holding it to unrealistic expectations of perfection that other fields do not meet.

Research Design & Methodology

The study conducted a limited review of literature published between January 1986 and May 2001 concerning the accuracy and reliability of screening and diagnostic tests in polygraph, medicine, and psychology, with 5,189 initial search hits produced, 1,158 articles and abstracts reviewed, and 145 found useful resulting in data on 198 studies. This systematic approach ensured comprehensive coverage of contemporary research across all three domains.

The study employed standardized psychometric measures to enable valid cross-discipline comparisons:

• Sensitivity: the ability to correctly identify true positives (deceptive individuals, diseased patients, psychologically impaired persons)
• Specificity: the ability to correctly identify true negatives (truthful individuals, healthy patients, psychologically normal persons)
• Kappa coefficient: measuring inter-rater agreement beyond chance, where 1.0 represents perfect agreement and 0.0 represents no agreement

Data collected on accuracy, agreement, number of subjects, and number of studies were entered into a spreadsheet and double-verified for accuracy, then used to sort studies and quantify summary measures. The analysis distinguished between diagnostic tests (focused on specific issues with evidence of a problem) and screening tests (multi-issue examinations without specific allegations).

Results & Key Findings

The findings revealed that polygraph performance aligned closely with established medical and psychological assessment tools, though with notable differences between diagnostic and screening applications.

Field Screening Assessment Accuracy:

• Sensitivity: polygraph .59, medical .79, psychological .74
• Specificity: polygraph .92, medical .83, psychological .72
• Polygraph showed notably lower sensitivity but higher specificity than medical and psychological screening tools

Inter-Rater Reliability (Kappa): Among readers in polygraph, medicine, and psychology kappa was .77, .56, and .79 respectively. Remarkably, polygraph examiners demonstrated stronger agreement than physicians and nearly matched psychologists.

The average accuracy reported for polygraph was similar to that reported for MRIs, CT scan, and ultrasound. More specifically, the accuracy of detecting deception by polygraph was similar to detection of breast cancer by ultrasound (.90) and more accurate than detection of breast cancer by x-ray (.80).

Reports in the literature of polygraph's accuracy and reliability (agreement) on specific issues appear to be consistent with published studies on medical and psychological assessment tools, though there is an enormous range of accuracy and agreement not only within polygraph but also medicine and psychology, and although there were very few polygraph screening studies, accuracy reports were lower than those in medicine and psychology.

Discussion & Significance

This comparative analysis fundamentally reframed the polygraph accuracy debate by establishing that diagnostic tests across professional fields exhibit substantial variability and imperfection. The study demonstrated that polygraph testing, particularly in diagnostic (event-specific) contexts, performs within the range of accuracy observed in widely accepted medical procedures. The high specificity (.92) of polygraph screening is particularly significant, indicating strong protection against false positive errors—falsely identifying truthful individuals as deceptive.

The lower sensitivity in screening contexts (.59) reflects a known trade-off in polygraph practice: conservative scoring standards that prioritize avoiding false accusations at the cost of missing some deceptive individuals. The summary statistics are not statements of accuracy for a particular procedure or profession, and there is much more that would need to be done to develop that level of precision, acknowledging the limitations of aggregated data.

The finding that polygraph examiners achieve higher inter-rater reliability than physicians challenges assumptions about subjective interpretation being a unique weakness of polygraph. The study suggests that the human-technology-examinee connection in polygraph mirrors diagnostic radiology, where expert interpretation of physiological data is standard practice and widely accepted despite inherent subjectivity.

Limitations & Considerations

The study faced significant constraints in available polygraph screening literature compared to the extensive medical and psychological databases. The vast majority of polygraph research focuses on diagnostic (event-specific) testing rather than screening applications, limiting the robustness of screening accuracy estimates. The 15-year literature window, while substantial, may not capture earlier foundational studies or the full evolution of polygraph techniques.

Aggregating diverse studies across multiple procedures, conditions, and populations inevitably obscures important variations. Medical diagnostic tools included ranged from sophisticated imaging technologies to simple screening questionnaires, while polygraph studies varied in examination formats, examiner training, and scoring methods. The study explicitly noted that summary statistics cannot be interpreted as definitive accuracy rates for any specific procedure or professional application.

Practical Applications

This research provides critical context for policymakers, attorneys, and practitioners evaluating polygraph evidence. Rather than expecting polygraph to achieve impossible perfection, stakeholders can assess it against realistic benchmarks set by medical and psychological diagnostics. The findings support the use of polygraph as an investigative tool and decision aid, comparable to other professional assessments that inform—but do not solely determine—consequential decisions.

For polygraph examiners, the study validates the scientific standing of their methods while highlighting the need for continued research, particularly in screening contexts where accuracy lags behind medical and psychological counterparts. The high inter-rater reliability underscores the value of standardized training and quality control. Consumers of polygraph services should understand that, like medical tests, polygraph results carry probabilistic rather than deterministic information, with accuracy varying based on application type, examiner expertise, and specific circumstances.